Guide complet 2026 pour utiliser l’intelligence artificielle dans son entreprise — Transcript
Full transcript
- 0:00This AI revolution proves that it's
- 0:01much more of a revolution in usage than
- 0:03in technology. 90%of French people use
- 0:05AI every day for two main use cases:
- 0:06researching on ChatGPT and writing
- 0:08emails. Those are exactly the two use
- 0:10cases that I advise against doing.
- 0:11There is an AI topic that is going to
- 0:12be huge. You won't be able to bill for
- 0:13a service the way you billed for it
- 0:14yesterday. That's over.
- 0:21The first thing when integrating AI is
- 0:22to avoid projects that are too big. We
- 0:24are missing out on 80%of what ChatGPT
- 0:26really is. And so we must start
- 0:28learning how to use all these tools
- 0:29properly. Start by identifying all your
- 0:31office tasks, do them with AI, and you
- 0:33will all save an hour and a half per
- 0:35day. Service companies are going to be
- 0:40transformed by AI. If you are
- 0:42well-equipped with AI, you either
- 0:43skyrocket your margins or you lower
- 0:44your prices compared to competitors and
- 0:46grow in volume. Today, we have a
- 0:47technology that makes people dream and
- 0:49promises incredible things, but I think
- 0:51they are completely mistaken. Thank you
- 0:56for joining us in ever-increasing
- 0:57numbers to listen to Le Déclic. If you
- 0:59haven't already, I invite you to
- 1:00subscribe, turn on notifications, and
- 1:02share a quick comment with us. It
- 1:04allows us to have more and more
- 1:05inspiring guests each time. And if you
- 1:07prefer listening to the audio format,
- 1:09know that it's also possible during
- 1:10your commute, work sessions, or even
- 1:12while exercising. You can find us on
- 1:15all listening platforms according to
- 1:17your preferences, as well as exclusive
- 1:18episodes or just the one you're
- 1:20currently listening to. Thanks again,
- 1:22it really allows us to keep evolving
- 1:25every time. And now, let's get on with
- 1:27today's episode. Hello and welcome to
- 1:28this new episode of the Le Déclic
- 1:30podcast. I am very well accompanied
- 1:31today, as always, with Martin Pavanello
- 1:33. How are you doing?
- 1:34I'm doing great, thank you. Thanks for
- 1:35the invitation.
- 1:36With great pleasure. Thanks for being
- 1:37here. We've been wanting to record this
- 1:39episode for a while. You have already
- 1:41had the opportunity to speak at one of
- 1:42our immersions on entrepreneurs.com.
- 1:44You have a superb background. There is
- 1:46an enormous amount to talk about
- 1:47regarding a subject that is currently
- 1:49very popular: artificial intelligence.
- 1:52We are going to talk about it at length
- 1:54through your background, through
- 1:55current events, and especially through
- 1:57how to implement it concretely in a
- 1:58company, whatever it may be. You have
- 2:01been the co-founder and CEO of Mysteria
- 2:03since October 2023, an agency
- 2:04specializing in generative AI
- 2:05integration. Your goal is to
- 2:08democratize the use of generative
- 2:10artificial intelligence for everyone,
- 2:12making AI a driver of competitiveness
- 2:14and transformation for professions and
- 2:16businesses. You have now trained over
- 2:1925,000 people in AI in France, with 250
- 2:21corporate training sessions per month
- 2:23and over 100 employees in just 2 years;
- 2:26that is quite exceptional. AI is
- 2:28undoubtedly the most popular topic, but
- 2:31perhaps also one of the most
- 2:33misunderstood today. We want to
- 2:36understand with you how to use it
- 2:37concretely in business and in life to
- 2:39make it a real performance lever.
- 2:41Welcome, Martin, to Le Déclic.
- 2:43Thank you very much, a great
- 2:44introduction. You've already said a lot
- 2:45of things.
- 2:45Thank you very much, my pleasure. And
- 2:47there is still a lot more to say,
- 2:48believe me. You were in real estate at
- 2:50the start, and in a few months, you
- 2:52became one of the faces of AI in France
- 2:54, particularly in real estate. But not
- 2:57only that—one of the most visible and
- 2:59well-known. What made you make this
- 3:01pivot?
- 3:02Yeah, that pivot—actually, I created
- 3:04my first company in real estate when I
- 3:06was in school, which is why I left
- 3:08school before getting my degree, and we
- 3:10were thinking about how to monetize the
- 3:12base we had created. In fact, we
- 3:14created a newsletter for real estate
- 3:16that addressed real estate agents and
- 3:17independent representatives. 20,000 to
- 3:2030,000 people followed it; we did a lot
- 3:22of advertising, so we monetized that
- 3:23way, and at one point, we wondered if
- 3:25there was another way to monetize. And
- 3:28actually, at that moment, the need to
- 3:29monetize the business differently
- 3:31intersected with the arrival of AI. So
- 3:34when I asked myself that question, it
- 3:35was two months before the release of
- 3:36ChatGPT. So you see, it was all quite
- 3:38linked. I had done a bit of AI in
- 3:39school; I did a business and
- 3:40engineering curriculum. So, the need to
- 3:42monetize, AI arriving, and secondly, we
- 3:45also felt that beyond AI arriving,
- 3:47there was a need to be supported
- 3:49because we saw everything going very,
- 3:51very fast. And the fourth point that
- 3:53tipped the scales was this need to
- 3:55build a slightly bigger company, if you
- 3:56will. I had this real estate media
- 3:59outlet; it was a great company that
- 4:01made a lot of money—and still exists
- 4:02today—but you can't scale it because
- 4:04we had 40 clients who were the 40
- 4:06companies in the ecosystem. Uh, so we
- 4:09were saying it's great, there are two
- 4:10of us, my partner and I, the company is
- 4:12doing well, we're making a very good
- 4:13living, but I can't turn it into a
- 4:15company that makes 10, 15, or 100
- 4:16million in revenue. And so, when you
- 4:18put these four ingredients together, we
- 4:20thought, well, the arrival of AI comes
- 4:21at just the right time. It reminds me
- 4:23of what I used to do in class. It
- 4:24really matches the need to perhaps
- 4:25train real estate agents, and so there,
- 4:27we can build a big company. So, all of
- 4:29that together. We really started in
- 4:30March '23, not as a formal business.
- 4:33That's why we tend to say October,
- 4:34because first we launched a newsletter
- 4:35about AI.
- 4:36thinking, well, we have this base of
- 4:3830,000 real estate agents, let's
- 4:39explain to them how they can use a bit
- 4:41of AI. So we...I originally created
- 4:43Misteria for that, which was a
- 4:44newsletter that talked a bit about news
- 4:45every week, and actually still exists
- 4:47today. And in fact, well, it
- 4:48overwhelmed us, because not only did
- 4:50those real estate agents join the
- 4:51newsletter, but many more people joined
- 4:53the newsletter. And so, we quickly grew
- 4:55to 40, 45,000 people on the AI
- 4:56newsletter, which meant that between
- 4:58March and September, we asked ourselves
- 4:59: "What do we do with this base we
- 5:01created around AI?" And in October, we
- 5:02turned it into a real company. But
- 5:04that's, in a way, how the transition
- 5:05happens: from "I have a real estate
- 5:06media" to, actually, I'm switching to
- 5:07an AI service company.
- 5:08We'll come back to the timeline of
- 5:10events and what followed. But before
- 5:12that, what made you dive into
- 5:13entrepreneurship, this adventure?
- 5:15school
- 5:15well, I also think it's a cocktail of
- 5:17many things. I was in business school,
- 5:19I was at HEC, and well, I have parents
- 5:20who are entrepreneurs, my brother is an
- 5:22entrepreneur, my grandparents were
- 5:23entrepreneurs. So, I'm quite convinced
- 5:25that intuitively and without really
- 5:27meaning to, you're a bit conditioned to
- 5:28do that, because you don't have parents
- 5:30who come home in the evening saying, "
- 5:32My boss did this, my boss did that."
- 5:33They talk more about their problems
- 5:35with their employees, about how they
- 5:36can sometimes go on vacation thanks to
- 5:38what they earn, about their
- 5:39entrepreneurial challenges. So all of
- 5:40that puts you in a mindset from the age
- 5:42of 10 where you think, "That looks like
- 5:43a pretty interesting job." So I think,
- 5:45subconsciously, I've wanted to do this
- 5:47for a long time, but I didn't want to
- 5:49skip any steps. In school, I did a
- 5:52classic prep course and a classic
- 5:54business school to get a standard
- 5:55management background. So, I really
- 5:58didn't want to rush things. But the
- 6:00real trigger—which fell on fertile
- 6:02ground—was during my gap year; I did
- 6:04a 6-month internship in real estate
- 6:06investment while at HEC and I loved it.
- 6:10You buy big assets, you resell them,
- 6:11you restructure them; it was quite
- 6:12fascinating, but I was an intern and I
- 6:14couldn't stand that. h
- 6:15In fact, you make slides all day, you
- 6:17work on Excel sheets all day, and you
- 6:18see your colleague next to you who is
- 6:2010 years older, and he's actually
- 6:21having fun. And I told myself, "Well, I
- 6:23want to do what he's doing; I don't
- 6:24want to be the one making the slides."
- 6:26So that's when I stopped at the end of
- 6:27my internship. I said, "I'm starting a
- 6:29company for the second half of my gap
- 6:30year, I'm going to try to build a
- 6:31business because that's what I want to
- 6:33do." And I want to have fun, I want to
- 6:34see clients, I want to be out in the
- 6:36field; I don't want to be in a suit
- 6:37behind my slides and never see anyone.
- 6:39So that’s how the click happened and
- 6:40that’s when I started the media. OK,
- 6:42clear. So you see this market
- 6:44opportunity, you see a need, and you
- 6:46create Mysteria A as a newsletter. It
- 6:48starts to take off. At that moment, AI
- 6:50wasn't necessarily as popular as it is
- 6:52today. I feel like we're talking about
- 6:5410 years ago, but in reality, it's only
- 6:55been 2 years. An enormous amount has
- 6:57been done. It’s a topic everyone is
- 6:58talking about. Where do you go to find
- 7:01information so you can quickly master
- 7:03the subject, be able to simplify it,
- 7:05pass it on, and, above all, help those
- 7:07who need to integrate it into their
- 7:09daily lives? That's a very good
- 7:13question because, as I said, I did a
- 7:15bit of AI in school; I studied AI in
- 7:17early 2022, so when I was still in the
- 7:19program, ChatGPT didn't exist. So, we
- 7:22talk about AI, we talk a bit about
- 7:23generative AI, as we mentioned earlier.
- 7:25Generative AI technology is much older
- 7:26than 2020, well, it's older than 2022,
- 7:28but there was no ChatGPT yet, so it
- 7:29hadn't been democratized. So what we
- 7:31see in school is quite theoretical,
- 7:32which is why that's not really what I
- 7:34highlighted later in the newsletter.
- 7:35And actually, my main source of
- 7:37information is the US. I subscribe to
- 7:40AI newsletters on Twitter, I follow the
- 7:4215 biggest AI creators, I've been doing
- 7:44that from the start. And I've already,
- 7:46just by telling yourself, by the way, I
- 7:47think for those who want to start
- 7:48companies, it's a great thing you can
- 7:50duplicate in any industry: tell
- 7:51yourself "I'll look at what's happening
- 7:52in the US" and. In France, we're all
- 7:54about 6 months behind, so it's a good
- 7:55business. So I say: "I'll look at what
- 7:57they're doing, I'll learn from what
- 7:58they're doing and I'll see if I can't
- 7:59do the same thing in France." So, I
- 8:01feed a lot, a lot on American content.
- 8:03And it's a big issue with AI, because
- 8:05you have many experts who have been
- 8:06telling us for 3 or 4 years: "Yes, but
- 8:08you come from real estate, so you know
- 8:10nothing,""Yes, but you became an AI
- 8:11expert even though you didn't go to
- 8:13engineering school for 5 years.""Yes,
- 8:15you don't know how to develop a machine
- 8:16learning tool, so you're not an AI
- 8:18expert." But I think they are
- 8:19completely wrong. In fact, I think this
- 8:21AI revolution, what it proves to us, is
- 8:22much more of a revolution in usage than
- 8:24in technology. For the tech, there are
- 8:26four companies in the world creating
- 8:27incredible models that are very complex
- 8:29, you know, requiring billions in
- 8:30investments. But that's an ecosystem of
- 8:3310,000, 20,000, 50,000 people worldwide
- 8:35. The users of ChatGPT are billions of
- 8:37people worldwide. You see, I personally
- 8:39decided I wanted to understand use
- 8:41cases and implications for the general
- 8:42public much more than claiming to be an
- 8:44AI expert who knows how to develop
- 8:45machine learning algorithms. So that, I
- 8:47don't know how to do. Uh
- 8:48And so that has really been—it's
- 8:49still my source of info, by the way—
- 8:51is to go and see a bit what companies
- 8:53are doing. So, drawing inspiration from
- 8:54the US and, secondly, seeing what
- 8:55companies are doing on the ground. My
- 8:57communication today, when you look at
- 8:58it, isn't about teaching people what
- 9:00machine learning is. It's about telling
- 9:01them, here is a bit of what's happening
- 9:02in the companies we support. Here are a
- 9:04few use cases for ChatGPT that you can
- 9:05use. Here are the few ChatGPT
- 9:07competitors that are interesting. Here
- 9:08are the processes that you can put in
- 9:10place. And that, actually, is just
- 9:11field experience and a bit of
- 9:12monitoring what's happening in the US.
- 9:14But that’s the big shift: AI went
- 9:16from being in the hands of experts for
- 9:1810 years to now being something very
- 9:19democratized. You just have to be a
- 9:21little interested in what companies are
- 9:22doing with it.
- 9:22Completely.
- 9:24And precisely to understand the
- 9:25revolution, do you have a few use cases
- 9:27like that to share with us that can
- 9:28resonate immediately with everyone
- 9:29listening?
- 9:30Yeah, absolutely. There are plenty. And
- 9:32the main takeaway is that we see this
- 9:33as a very office-based revolution.
- 9:35That’s also what needs to be
- 9:36understood. AI is first and foremost an
- 9:38office revolution. So today, all your
- 9:40office tasks are being transformed by
- 9:42AI. The task of making slides. I make
- 9:44slides today using a tool called James
- 9:46Spark. It’s a tool that makes slides
- 9:48in the Boston Consulting Group style in
- 9:503 minutes. The guys, the guys at BCG,
- 9:52my classmates who are at BCG, I think
- 9:54when they see that, they pull their
- 9:55hair out. You can use the same colors,
- 9:57the same logos, you get the same
- 9:58magnificent charts. Except the info
- 9:59comes from an AI that did it for you in
- 10:012 minutes. So, slide tools are quite,
- 10:03quite incredible. Meeting summary tools
- 10:05, and here too, these are very simple
- 10:06things. But I use Nouta a lot. At first
- 10:09, you tell yourself, well, do I
- 10:10actually make meeting summaries? I
- 10:11never did them. I never did meeting
- 10:13minutes. So I always told myself it’s
- 10:14not necessarily a use case for me.
- 10:15Actually, I realize now that it has
- 10:17applications much further than just
- 10:19making a meeting summary. I’ll give
- 10:21you an example. In every video call I
- 10:22have, Nouta is my assistant; it comes
- 10:24with me into the call. At the end, I
- 10:26don’t look at the report. Well, the
- 10:27report, I don’t even look at it.
- 10:28However, two days later, I have a
- 10:30commercial proposal to make. I take
- 10:31Nuta's transcript, I take their report,
- 10:33I send it to my slide tool, and it
- 10:34makes 10 slides at once, you see. h
- 10:37And here, you go from a tool that just
- 10:38makes meeting content to one that
- 10:40actually helps you create a highly
- 10:42personalized sales proposal. Well, it
- 10:44makes a slide for me saying, "Hey,
- 10:45watch out, the client mentioned their
- 10:47training need is more advanced, not
- 10:49really for beginners," and so on; so
- 10:50that's a tool, for example, that I use
- 10:52every day now. So everything regarding
- 10:53slides, everything regarding reports,
- 10:55and then standard ChatGPT. I use
- 10:57ChatGPT a lot for myself; I use it for
- 10:59questions here, perhaps to change up
- 11:00what we’re used to. Everyone tells
- 11:02you it helps you do all the
- 11:03low-value-added tasks. It’s true,
- 11:05there are plenty of low-value-added
- 11:06tasks where AI is very strong. It’s
- 11:08also very strong for high-value-added
- 11:10tasks. I have the latest model, GPT-5
- 11:12Pro. When I'm recruiting, I use it for
- 11:15brainstorming. Meaning, I’ll tell it:
- 11:17"I’m hiring a salesperson, here is my
- 11:19revenue, here is the margin I want to
- 11:21make, here is the number of calls they
- 11:22can make per day, here is the type of
- 11:24client I’m going to put in their
- 11:25hands, etc.""Give me the revenue
- 11:27targets I should give them to close
- 11:28deals.""And then make me a sensitivity
- 11:30table based on what they sign, and how
- 11:31much margin I’ll make." It’s a job
- 11:33that would take me hours to put on
- 11:35paper and so on. And here, it reasons
- 11:37for 5 minutes, where it really helps me
- 11:39with a strategic question. So you can
- 11:41really combine the two: quick wins,
- 11:42fast things, but also strategic
- 11:44reflections where AI becomes very, very
- 11:45strong.
- 11:46And this ability to use AI daily,
- 11:49creating a new reflex, is a real asset.
- 11:54And that is why many, ultimately, don't
- 11:55realize how much it can change an
- 11:56entrepreneur's life, whether it's for
- 11:58handling emails, office work, or using
- 11:59it in recruitment. But we could—we
- 12:01could—we could mention it for every
- 12:03single point. You could even tell
- 12:05yourself, "I'll take this breakthrough
- 12:07episode, push it into the AI, get a
- 12:08transcript, a concrete summary of all
- 12:10the clear actions I need to implement
- 12:11in my business, and I'll do that every
- 12:13time for every episode." Uh, do you
- 12:17have a trick, a specific point that you
- 12:19tend to share, even within the company
- 12:21or with people you train, for
- 12:22integrating artificial intelligence as
- 12:25a habit? The first thing when
- 12:28integrating AI for me is to avoid
- 12:30projects that are too big. Actually,
- 12:32you end up missing out on 80%of what
- 12:34ChatGPT is, for example. There are many
- 12:36companies that tell me, "Okay, we've
- 12:38deployed 100,000 euros on a project,
- 12:40we're going to try to develop a tool
- 12:41that will automate 100%of our processes
- 12:43." I say, "Do your employees even know
- 12:45how to use ChatGPT or Claude or an
- 12:46equivalent yet?" No. So let's start
- 12:48with that. That's a quick win. It’s
- 12:50actually when you train quickly and get
- 12:52a few licenses that you can already
- 12:53achieve major productivity gains. So we
- 12:55really insist on that. There are plenty
- 12:58of things to go after, as the Americans
- 12:59say, "low-hanging fruit," you see.
- 13:01There are lots of easy things to go
- 13:02after like that. And so you have to
- 13:04start by really learning how to use all
- 13:06these tools. So I would advise them to
- 13:08do that instead. Start by identifying
- 13:10all your office tasks, doing them with
- 13:12AI, and right there, you'll all save an
- 13:14hour and a half a day. That is really
- 13:17the big topic. And after that, we'll
- 13:19get into more advanced automation. Are
- 13:20there processes we can automate? Are
- 13:22there customer services where we can
- 13:23add more AI, where we go a bit further,
- 13:25and so on? But in any case, I think the
- 13:26biggest part of integrating AI is not
- 13:28to go too fast. There's an MIT study
- 13:31that came out two weeks ago saying that
- 13:3395%of AI projects don't finish and have
- 13:35no ROI for companies. And when you dig
- 13:38deeper—that's the headline on TF1,
- 13:39published everywhere, Les Échos,
- 13:41everyone is happy. But actually, when
- 13:42you dig deeper, you realize that all
- 13:44the projects mentioned were, first of
- 13:45all, projects done by large
- 13:46corporations. It's all development and
- 13:48automation projects and AI agents.
- 13:51These three things are not yet mature.
- 13:53And likewise, when you ask those three
- 13:54companies, have you trained your
- 13:56employees on Copilot, GPT, or Claude?
- 13:57No. So today, we have technology that
- 13:59makes us dream, promising us incredible
- 14:01things, which are sometimes a bit
- 14:02disappointing because, well, the agents
- 14:04aren't very mature. So there's no point
- 14:06in wasting too much time and money on
- 14:07that, which makes us miss out on what
- 14:09SMEs are succeeding at. SMEs are
- 14:10actually succeeding very, very well at
- 14:12mastering one or two AI tools perfectly
- 14:14, and you'll already save a crazy
- 14:15amount of time, you know. So that would
- 14:18be my recommendation on how to approach
- 14:20, or rather, what steps to take to
- 14:22implement AI. I would say don't look
- 14:25for a miracle cure, but rather go learn
- 14:28how to use it. Experts say we only use
- 14:3015-20%of Chat GPT's capacity.
- 14:32Completely.
- 14:33It's a shame. It's still a tool that we
- 14:36struggle to grasp. I think I saw that
- 14:3990%of the French people who use AI
- 14:41every day have two main use cases.
- 14:43Doing research on Chat GPT and writing
- 14:44emails. Those are exactly the two use
- 14:46cases that I advise against. Well,
- 14:47emails maybe, but doing research, no.
- 14:50So you see, we're still there. We have
- 14:51the Google reflex, it's a tool we don't
- 14:53know well enough, but honestly, they
- 14:54didn't develop this tool and they
- 14:56didn't just raise 500 billion recently
- 14:58so we could write emails and do
- 14:59searches with it. Otherwise, there's
- 15:01Google. So it's meant to go way, way
- 15:02beyond being just a simple tool. So I
- 15:04would recommend that: few tools, find
- 15:06concrete use cases, and already handle
- 15:08that office work before moving on to
- 15:09larger projects.
- 15:10Very clear. We'll talk more about usage
- 15:12and concrete cases in a moment. But
- 15:14before that, let's go back to the
- 15:15timeline of Mysteria. So October 2023,
- 15:18you launch Misteria after doing this
- 15:21newsletter. Uh, what are the first
- 15:23steps? How does it happen? Because
- 15:25quickly you have some success, quickly
- 15:26it scales quite hard, you're going to
- 15:28have to staff up, you're going to have
- 15:29to structure. Uh, we're also going to
- 15:32talk about revenue and results, the
- 15:33number of clients you've been able to
- 15:34support. It's pretty crazy. And then
- 15:36there’s also the door left open at
- 15:38some point to, I imagine, the siren
- 15:39song, because everyone is asking for AI
- 15:40. At first, you start in real estate,
- 15:43and then, well, you readjust. It might
- 15:45be interesting for you to tell us that
- 15:46story.
- 15:46Yeah, absolutely. So, we launched in
- 15:49October 2023; actually, during the
- 15:50summer, we realized too many requests
- 15:52were coming in from the newsletter for
- 15:53training and consulting. We weren't
- 15:55answering them, so there was a need, we
- 15:56had to go for it. So we created those
- 15:58two services. We hired quite quickly.
- 15:59We hired two or three people in October
- 16:02, mostly consultants. So we raised a
- 16:04bit of money. If you look online, we
- 16:05raised a little bit of money from
- 16:07business angels because, well,
- 16:09consultants coming out of top business
- 16:10schools cost you quite a bit—several
- 16:12tens of thousands of euros a year,
- 16:14sorry, a month, all included. Plus your
- 16:16clients, the projects last 2 or 3
- 16:18months, so they pay you after 4 months.
- 16:20So, in short, you have a working
- 16:20capital requirement of a few months. So
- 16:22we figured we needed to hire a bit
- 16:23right away, so we raised a little money
- 16:24.
- 16:25How much did you raise?
- 16:26We raised 800,000 euros.
- 16:27OK. We had a little less than a million
- 16:29euros, we had said a million because
- 16:30there was a mix of equity and debt, but
- 16:32we didn't end up taking debt, so a bit
- 16:33less in equity. The company gave up a
- 16:36little over 10%of its capital, fairly
- 16:38standard for a round with business
- 16:40angels. So we did that round, we hired
- 16:42the first three or four people. It was
- 16:45a bit of a struggle at first. As you
- 16:47said, we started a lot in real estate.
- 16:48We have a pretty strong real estate
- 16:49network with my partner because we have
- 16:51this company together and he had
- 16:52another company too. So it's my brother
- 16:54Vincent who is my partner. He has
- 16:55another real estate company that does
- 16:57consulting, which, for its part,
- 16:58doesn't target real estate agents but
- 16:59large real estate firms. So we started
- 17:01by going to see them at first. We
- 17:02looked through that real estate lens.
- 17:04So we started with that. As you said
- 17:06earlier, it's an ecosystem that isn't
- 17:07mature at all. You arrive at a moment
- 17:09where everyone wants AI, but when you
- 17:10start saying, "OK, it’s going to cost
- 17:12you X amount just to lift a finger."
- 17:14Everyone says, "Actually, I don't know
- 17:15what I need." I want to do some AI, but
- 17:16I don't know what. I don't want to use
- 17:17ChatGPT. GPT is too dangerous. Get
- 17:19training, okay, but on what tool?
- 17:20Copilot didn't exist. Develop a tool?
- 17:22Well, no, I don't want to put €
- 17:2450,000 on the table. So there it is, we
- 17:25arrive at the right time, it's the very
- 17:27beginning, but a time when companies
- 17:28aren't ready. So for a few months, we
- 17:31do some small training sessions, we do
- 17:33some small audit missions fairly
- 17:34quickly. It doesn't move much for 6
- 17:37months. I would say that for 6 months,
- 17:39we recruit, well it doesn't move much,
- 17:41we aren't recruiting, we're doing our
- 17:42little missions on our side, and we
- 17:44have a first shift that arrives in June
- 17:45'24 with a first training course that
- 17:47comes out for real estate agents where,
- 17:49in fact, immediately, we have thousands
- 17:50of people who buy the training, and
- 17:52there, we find ourselves with the need
- 17:53to staff trainers and recruit trainers.
- 17:55That was the first big step, I'd say,
- 17:57for Misteria, to have to say, actually,
- 17:59I need 10 or 15 trainers all at once.
- 18:02So there, that's a bit in June '24. We
- 18:04had this first momentum of recruiting,
- 18:06and with a model that quickly turned
- 18:08toward freelancers because, in fact,
- 18:10you are in a business where, even today
- 18:12if you take the 100 people, there are
- 18:1450 employees and 50 freelancers,
- 18:15because doing corporate training is a
- 18:17freelance business. In fact, when you
- 18:19have employees, and you tell Louis, "
- 18:20You are going to go train in Toulouse
- 18:21tomorrow," and Arthur, you tell him, "
- 18:23You are paid the same salary, but
- 18:24tomorrow you are going to go to Paris."
- 18:25Well, actually, they look at you at the
- 18:27end of the day and say, "Why me
- 18:27Toulouse and him Paris?"
- 18:28And that is not manageable. Everyone
- 18:30tells you, "Ah, freelancers must cost
- 18:32less.""No, the freelancer costs you
- 18:34more in daily rate, but you have a lot
- 18:35less hassle afterward, because in fact,
- 18:37with a freelancer, you say: 'I have a
- 18:38training in Toulouse, who wants to go?'
- 18:40And you always have someone who says: '
- 18:41Well, I'm not far, I'm up for it, I
- 18:43like it, I know this ecosystem well,
- 18:44etc.'"» So very quickly, we told
- 18:46ourselves we needed a lot of
- 18:47freelancers to provide our training
- 18:48services and a little bit of consulting
- 18:50too; for consulting, we are using
- 18:51freelancers more and more. So, the
- 18:53company was structured like that pretty
- 18:55quickly, and I would say with real
- 18:56momentum starting in June when we
- 18:58already integrated 15 to 20 employees
- 19:00in June '24; that was the first wave,
- 19:02and there, Mysteria went from 6 or 7
- 19:03employees to quickly 30 or 35 all of a
- 19:05sudden like that.
- 19:06That was a bit of a first step. And so,
- 19:08right away with this desire to
- 19:10structure it more or less 50/50 between
- 19:11freelancers and internal staff.
- 19:13Yes, quite quickly. Yeah.
- 19:14What functions did you keep in-house?
- 19:16And we understand that freelancers are
- 19:18mainly for the training and support
- 19:19side.
- 19:20Yeah, exactly. Internally, you keep
- 19:22what we call support functions. HR part
- 19:24. The HR part in a training company is
- 19:26essential because, in fact, you are
- 19:27staffing all day long; we were talking
- 19:28about 250 training sessions. Well,
- 19:30every month, you have to find 250 days
- 19:33available from trainers. So you are
- 19:35staffing everywhere left and right. So
- 19:36you need a strong team. Uh, that's
- 19:38important. And the same goes for the
- 19:40pedagogy side; in fact, you have to
- 19:41prepare the training. You have to
- 19:43provide feedback to the companies. So
- 19:46you need to have support for that. So
- 19:48pedagogy, HR, uh CFO and finance, we
- 19:50also need that because there are many
- 19:52small line items and so on, so we need
- 19:55to have fairly strong financial teams.
- 19:57After that, we have all the partners
- 19:59in-house. Partner/manager, these are
- 20:01all the people who will manage the
- 20:02consulting assignments. We have four of
- 20:04them today. They see 100%of the
- 20:05consulting assignments, they
- 20:06participate in 100%of the consulting
- 20:08assignments. In fact, when you do
- 20:09consulting, you still have these needs
- 20:11in a company for long-term support, and
- 20:12your client doesn't want to talk only
- 20:14to a freelancer, and I totally
- 20:15understand that. So, I have a partner
- 20:17who will manage the client, uh, the
- 20:19client Vinci. The partner will manage
- 20:22Vinci, and when they need to, they will
- 20:23put a freelancer on a training session,
- 20:24a freelancer on a bit of consulting,
- 20:26etc., but they are the one who follows
- 20:27the whole mission, who sees the client
- 20:28regularly, and so on. So our four or
- 20:31five partners are in-house, and then
- 20:32there are sales reps, a few sales reps,
- 20:33including on the B side as well.
- 20:35Excuse me, I interrupted you when you
- 20:36said "partner." Uh, are they associates
- 20:38? Are they associates at Misteria or
- 20:39not? or standard start-up stock options
- 20:41, but there are two main partners at
- 20:43Misteria, Vincent and me. So my partner
- 20:45and I, my brother and I, we own most of
- 20:47the com—the majority of the company,
- 20:49and then there are stock options for,
- 20:50indeed, what we call partners and some
- 20:52of the management positions at our firm
- 20:53.
- 20:55So it will soon be almost two years to
- 20:57the day since it was officially
- 20:59launched, beyond the newsletter. Uh, in
- 21:02terms of revenue over two years and
- 21:04growth, uh, where do you stand? So it's
- 21:07a company that made € 100,000 in
- 21:09revenue its first year. Well, the first
- 21:11year is September to December, really,
- 21:12or October to December, even. Uh, since
- 21:14we started in October, last year we
- 21:16closed at around 344k in revenue. This
- 21:18year, we should do a little over 7
- 21:19million in revenue with a profitable
- 21:21business model. So, uh, it's hard
- 21:23because doubling growth while
- 21:24maintaining a margin in B2B like that
- 21:26is quite difficult. But the company is
- 21:29very profitable. As a service company,
- 21:31you generally make between 15 and 30%in
- 21:33EBITDA margin. So we find ourselves
- 21:35somewhere between those two figures. Uh
- 21:36, so, uh, rather good growth, uh,
- 21:38prices that are trending downwards,
- 21:40which is also increasingly interesting.
- 21:42You see, the market is becoming
- 21:43structured, competitors are arriving,
- 21:45so you can't sell your training day for
- 21:46€ 2,500 like you did at the beginning
- 21:48. So we are pushing prices down a bit;
- 21:50I think that's fairly standard. But, uh
- 21:53, yeah, that's how it is today. The
- 21:55goals are quite high, you know. It's a
- 21:56company we want to take very, very far.
- 21:58Uh, the goal is to reach 100 million in
- 21:59revenue by 2026. That's pretty, pretty
- 22:01clear for us. And then, after that, to
- 22:03go further, do some M&A, we're starting
- 22:04to think about international expansion.
- 22:06Because we are training in eight
- 22:07countries now.
- 22:08Uh, so, rather very small teams. I have
- 22:10clients, uh, clients like Clarence, for
- 22:12example, who is a big client of ours,
- 22:14who tells us, "Well, it's the France
- 22:15team we're training, and now we're also
- 22:17going to train Asia, China—Asia, the
- 22:19Middle East, and the US." So for now,
- 22:21it's mostly our French clients taking
- 22:23us abroad, but we clearly want to
- 22:24expand internationally too. So, that's
- 22:28where we're at today, segmented into 70
- 22:31%B2B and 30%B2C. We also do some B2C
- 22:35training. More and more. So, that's how
- 22:38we're structured, and I think the
- 22:40company is destined to be bigger in B2B
- 22:42than B2C, even if we remain very humble
- 22:45in our approach. The ecosystem is so
- 22:47new that we really have no idea what
- 22:49the company will look like in five
- 22:51years. Will our clients need more
- 22:53training or more consulting? Today it's
- 22:5550/50, but what it will look like in
- 22:57five years is really hard to say. So,
- 22:59that's the current state of affairs.
- 23:01But trying to predict where we'll be in
- 23:03five years is very difficult.
- 23:04Absolutely. There are so many things
- 23:06you just mentioned that I'd like to
- 23:08touch on. You mentioned, quite rightly,
- 23:10that it's challenging to double your
- 23:12growth from one year to the next. Even
- 23:15early on, from 100k to 4 million. So I
- 23:18imagine it involved a lot of disruption
- 23:21, tough decisions, and needing to
- 23:23structure quickly without making
- 23:26mistakes. What have been the biggest
- 23:29hyper-growth challenges you've faced in
- 23:30such a short time so far?
- 23:32Yeah, that's a good question. The
- 23:34biggest growth challenge is structuring
- 23:36, that's for sure. The thing is, I'm 27
- 23:39, I've never had a CFO, and I've never
- 23:41done the HR or CFO job myself. So at
- 23:43some point, you have to trust people
- 23:45whose jobs you don't actually know.
- 23:46That's hard. You're forced to think,
- 23:48well, I hope the CFO is good and the HR
- 23:49manager is strong. It's been tough to
- 23:51accept that you no longer control
- 23:53everything happening in the company;
- 23:54even if you try to follow it all, you
- 23:56have to trust people. That was the
- 23:57first main challenge. The second
- 23:59challenge is the business model. Maybe
- 24:01it's a mystery, I don't know. We face a
- 24:03major business model challenge,
- 24:05especially regarding consulting. It's
- 24:07very hard to build a consulting firm
- 24:09that scales a bit like we do, because
- 24:11everyone talks to you about daily rates
- 24:13in consulting, how much you pay your
- 24:15guy, how many days you manage to place
- 24:16them, and so on. That works when your
- 24:19company has plenty of clients, like BNP
- 24:21spending 1 million euros with you every
- 24:23year for 10 years. So you know exactly
- 24:25where you're placing your consultants
- 24:26and so on. When you have clients who
- 24:28spend € 7,000 on consulting, then the
- 24:30next one spends € 50,000, and one
- 24:32month you have two consulting projects,
- 24:33and the next month you have 13,
- 24:34that’s when it becomes a nightmare. H
- 24:37So, in the beginning, we, all our...all
- 24:38our projects were done only by
- 24:40employees, only salaried consultants.
- 24:41And we had a major business model shift
- 24:43, telling ourselves: "Actually, this
- 24:44isn't sustainable." With the growth
- 24:45we're having, we won't be able to hire
- 24:47employees. Because, likewise, you sign
- 24:48three times as many projects in a month
- 24:50, so what do you think? Should I hire
- 24:51two consultants? Except that when these
- 24:53consultants arrive, well...you find
- 24:55them, and so on, and the projects might
- 24:57already be finished. So, in fact,
- 24:58potentially, they won't be working for
- 25:002 months. Except that you're hiring
- 25:01good profiles, so it costs a lot, so
- 25:03that's really a struggle. So that’s a
- 25:05big challenge: figuring out how to
- 25:06pivot our business model to absorb the
- 25:08growth of the Consulting BU, in
- 25:10particular. Training is easier. For
- 25:11training,
- 25:13tomorrow, if I launch a trainer
- 25:14recruitment drive in my newsletter, I
- 25:16get about 200 people applying every
- 25:17time. So when there's a need, you see,
- 25:20we can onboard 15. I think if we feel
- 25:23that we're selling a lot in September,
- 25:24in October I'll onboard 15. That,
- 25:26quote-unquote, is a bit easier. In
- 25:28consulting, it's much harder. So we
- 25:30shifted; we went from a model where all
- 25:32training and all projects were done
- 25:34exclusively by employees. to a model
- 25:36where all projects are done exclusively
- 25:38by freelancers who are managed by
- 25:40employees. But the guy who manages a
- 25:42project can manage 5, he can manage 10,
- 25:43and if he ever has to manage 15 in a
- 25:45month, he'll sleep less, but he'll be
- 25:46able to do it.
- 25:47because behind the scenes he uses
- 25:49freelancers who help him out. But guys
- 25:50who produce 5, 10, or 15 in a month,
- 25:52well no, normally when you're
- 25:53delivering on a project, you can't do
- 25:55more than is reasonable. So that was
- 25:58the big challenge we had, I would say,
- 26:01managing to really, really pivot the
- 26:03business model. Then, there's a
- 26:06financial challenge I didn't suspect
- 26:08either. It's one thing to run a company
- 26:11that makes 100k, 200k, a million, or 2
- 26:13million. At 7, thank goodness there's a
- 26:15CFO. It's a good thing there's a CFO.
- 26:16Especially in a business where when you
- 26:18have 7 million in revenue and sell
- 26:20projects for € 300k, it’s, quote
- 26:22unquote, fine, you can keep up. We sell
- 26:24daily rates of 2,000. Uh, well now we
- 26:26have bigger, much bigger tickets,
- 26:28it’s improving, but when we were even
- 26:30doing 3, 4, 5 million in revenue, they
- 26:32were all tickets between € 500 and
- 26:332,000. Uh.
- 26:34So you're managing a huge number of
- 26:35small line items. Plus, we have a part
- 26:37of the business that is state-funded.
- 26:39People can get their training
- 26:41reimbursed. So it's endless paperwork.
- 26:44So financially, you get to the end of
- 26:45the year and you say, "Oh wait, I
- 26:46actually have massive budget
- 26:47discrepancies." So we had to reconcile
- 26:49everything, and so on. But that’s a
- 26:50challenge where, if I start a company
- 26:52again tomorrow, the CFO aspect is
- 26:54something to take seriously. Now, it
- 26:56depends on the business, but it really
- 26:58must be taken seriously. Uh, for us,
- 27:00that was the issue. So, there you go,
- 27:02I’d say financial, business model,
- 27:03and sales too, because when you sell
- 27:05like we do, you aren't selling a
- 27:06product. You aren't selling a product,
- 27:08and, uh, I'd have a 4th point, actually
- 27:10, I'll mention it, but you aren't
- 27:11selling a product. So every morning you
- 27:13wake up and think, if I want to sign
- 27:15500k or 600k in services this month, I
- 27:17have to go find X number of new clients
- 27:20. That's hard when you're selling a
- 27:22subscription product. In December, when
- 27:24you wake up on December 1st, you
- 27:25already know the minimum revenue you're
- 27:27going to make that month. When we wake
- 27:29up, we don't know the revenue we're
- 27:30going to make for the month. So that's
- 27:31tough. And also, when the company grows
- 27:33slowly, it's fine. When it grows super
- 27:35fast, it's hard because regarding your
- 27:37investment choices—and this ties into
- 27:39my fourth point—the big, big, big
- 27:41challenge is building a profitable
- 27:42company while growing. Your investment
- 27:45choices have to be calculated down to
- 27:46the millimeter. You ask yourself, "
- 27:48Should I spend 5,000 euros here or not?
- 27:49" Because we absolutely want to build a
- 27:51company that makes a 15, 20, or 25%
- 27:52margin. That is really something we
- 27:54want to maintain. We don't want to say,
- 27:55"Okay, we accept making 5%or we accept
- 27:57losing money for the sake of growth."
- 27:59For now, we don't really want that. So
- 28:00balancing the two is hard. It means
- 28:02being really tight on all your
- 28:04financial line items. Why did we pay
- 28:06that much for this trade show? Why
- 28:07shouldn't we pay less for rent? All our
- 28:09teams negotiate all day long. I tell
- 28:10them: if you're buying something, you
- 28:12negotiate. Because actually, if we all
- 28:13negotiate 10%off, that's 10%less in
- 28:15expenses. For a company with our
- 28:16revenue, that saves hundreds of
- 28:18thousands of euros a year, even if it
- 28:19comes from small 10%cuts here and there
- 28:21on subscriptions, etc. So that has been
- 28:23—and still is—a big challenge. It's
- 28:25about how we manage to grow while
- 28:27keeping our employees in the mindset
- 28:29that while we are a tech startup to
- 28:31some, in terms of structure, we are
- 28:34actually an SMB. So an SMB has to
- 28:36invest 1 euro, and that 1 invested euro
- 28:38must bring back 2, 3, 4, or 5, but we
- 28:40can't just say, "Let's recruit
- 28:41infinitely." You see, with hiring,
- 28:43sometimes your sales director comes to
- 28:44see you and says, "Yeah, I want to hire
- 28:46two or three salespeople.""I think it
- 28:47would be good." Oh sure, if I had my
- 28:49investors 'money, okay. If I'm spending
- 28:51money that isn't mine, why not? But
- 28:53here, I tell him, "No, tell me why you
- 28:54want one, what is it going to bring you
- 28:56?""What is it going to cost us?"
- 28:57That’s a great challenge
- 28:59for us and for the managers because we
- 29:00spend a bit more time on this, but the
- 29:02manager is less experienced, our guys
- 29:03are young and all. So that’s a good
- 29:05challenge. How do I grow, how do I not
- 29:08spend at any cost in order to stay
- 29:10profitable?
- 29:10It’s a real topic, optimizing growth
- 29:12and EBITDA in the end.
- 29:14People often say it's not necessarily
- 29:15possible to do both, and the challenge
- 29:16here is to succeed in doing both as
- 29:18much as possible.
- 29:18Yeah, exactly. I know a lot of people
- 29:21say you need, I think it's 30%, they
- 29:23tell you that if your EBITDA margin
- 29:25plus your growth is over 30%, your
- 29:27service company is in very good health.
- 29:30So we're well above that, so that's
- 29:32good for now. I think that's it, 30 or
- 29:3440%. So you see, we're trying, and
- 29:36we'll get there. We won't have 100%
- 29:37growth every year, and we won't have 25
- 29:39%margins every year either. We're quite
- 29:41realistic. The more CAC 40 companies
- 29:42you take on as clients, the more you'll
- 29:44get squeezed on price, trade-offs, and
- 29:46so on. So we'll end up like a classic
- 29:48consulting or training firm, hovering
- 29:50between 15 and 20%margin. That's pretty
- 29:51much certain. So that means we have to
- 29:53keep growing by 10, 15, or 20%. But
- 29:54yeah, it's very, very hard. Very, very
- 29:56hard. And it really stands out. No, no,
- 29:58I talk about it a lot, but that's where
- 30:00the French Tech scene is good. I love
- 30:02the French Tech scene, it's cool and
- 30:04all, but building a tech startup where
- 30:05you raise 100 million is so different
- 30:07from building a body shop or a bakery
- 30:09where you only invest a euro once
- 30:10you've actually earned it. Obviously, I
- 30:13respect them, but I admire the baker so
- 30:15much more, who has to pay salaries and
- 30:17thinks: "How do I manage to open three
- 30:19bakeries?" Well, I have to make sure my
- 30:21EBITDA margin on bakery number one is
- 30:23higher so I can invest, but what I'm
- 30:24investing is a risk. So that's what
- 30:26fascinates me, and we want to run our
- 30:28company like that. So that is often
- 30:30underestimated. Everyone says, "Oh
- 30:32great, they raised 100 million, they're
- 30:34so strong, look, they have huge growth.
- 30:36" I always say growth is one thing, but
- 30:37at what price? Growing with other
- 30:39people's money is not the same as
- 30:40growing with the money you generate
- 30:42yourself, and you have a business model
- 30:43quite similar to mine. So, you see what
- 30:45I mean? That's quite difficult. So it's
- 30:47the biggest challenge.
- 30:49Completely. What makes you take this
- 30:51decision to want to scale the company
- 30:53this way? Because, ultimately, uh,
- 30:55there's a culture that's quite, uh,
- 30:56quite young, quite tech-oriented, quite
- 30:58US-style, quite start-up overall,
- 31:00whether it's the AI, the average age of
- 31:02the employees, the hyperscale at
- 31:03certain moments, the fact that at the
- 31:05start you raise 100 million euros—
- 31:07that's not trivial either, it's
- 31:09strategic, 1 million euros but on a
- 31:10valuation of 10, excuse me—it's, it's
- 31:12, it's strategic, it's well thought out
- 31:14. You could have also told yourselves "
- 31:16Okay, well, we're going to keep wanting
- 31:17to do that to flood the market and ride
- 31:18the wave." However, you have a
- 31:21philosophy that I completely share,
- 31:23which is to invest every euro
- 31:25intelligently and build something solid
- 31:27, even with major growth, much like the
- 31:29small businesses you mention.
- 31:33There are plenty of reasons, I think.
- 31:34The first is a bit about the market
- 31:35structure, actually. I think in a world
- 31:37where VC is doing much worse in France,
- 31:38you don't raise money like that anymore
- 31:39. Now, we have a specific theme, so
- 31:41that helps, all the same. But it's
- 31:43harder to raise VC money today, so we
- 31:44arrived at that point, and we said to
- 31:46ourselves it's not necessarily the
- 31:47track we want to follow. We sell
- 31:49services. There are few service
- 31:50companies today that get VC funding.
- 31:52Hm.
- 31:52So, uh, we told ourselves, look, we're
- 31:55in our place, we make money from day
- 31:57one, so we might as well stay the
- 31:58course. Yeah, that was, that was pretty
- 32:01much it at the start, and besides, that
- 32:03doesn't stop you from growing. And, by
- 32:05the way, we see all the companies that
- 32:06have experienced growth. I have many,
- 32:07many friends, and especially Vincent,
- 32:09my partner, who is a bit older. He grew
- 32:11up, he was 25 or 24 years old, between
- 32:1320 and 25, right in the middle of the
- 32:15French Tech boom. Uh, I was that age
- 32:18when French Tech was already doing a
- 32:19little worse. So he had, uh, 50 friends
- 32:21who started companies and all raised 10
- 32:23or 15 million euros. Not a single one
- 32:24of them ended up taking any cash home.
- 32:25Either the companies crashed, or they
- 32:27ended up having an exit below their
- 32:28previous funding rounds, so they didn't
- 32:29make any money, and so on. Not for
- 32:31everyone, obviously, but for many. So,
- 32:33I, I kind of grew up with that too.
- 32:35That didn't excite me, the idea of
- 32:37saying, "Okay, I'm going to raise and
- 32:38raise and raise, and I won't touch a
- 32:39single dime for a while." Uh, I'd need
- 32:43the macro and the economic climate to
- 32:45be good in 5 or 10 years if I want to
- 32:47make money by selling. I told myself,
- 32:49that excites me. And then, when I said
- 32:51everyone in my family was a bit of an
- 32:53entrepreneur, it was always with very
- 32:54classic SMEs. So we also told ourselves
- 32:56, look, this is a healthy business,
- 32:57there are a lot of things to do in AI
- 32:59here. That doesn't stop us from growing
- 33:01, it doesn't stop us from building a
- 33:02company; One Point is worth a billion,
- 33:03you see, our competitor, in quotes—I
- 33:05say quotes because they are much bigger
- 33:06than us and have existed for 20 years.
- 33:08But they are an example for us.
- 33:09Artefact, One Point, these are AI
- 33:11consulting firms that are very big.
- 33:13Well, they've been around for 10 or 20
- 33:15years and the companies are worth
- 33:16billions, so if there's a path to do it
- 33:18that way while making money along the
- 33:20way, well, that's what interested us,
- 33:22and we're going to continue down that
- 33:23path. Now, when you're making money,
- 33:26you don't raise from VCs anymore. But
- 33:27there you go. And another point,
- 33:29perhaps, is that I don't know if we're
- 33:31in an ecosystem where raising 10
- 33:33million euros would really help us grow
- 33:35. You see, when you sell a product or a
- 33:38tool, you go all-in on communication,
- 33:41you hire a bunch of salespeople to make
- 33:43calls all day, and you grow fast. For
- 33:46us, it's so much of an execution-based
- 33:48business—where you need strong
- 33:50execution to be effective—that if you
- 33:51gave me 5 million euros tomorrow, I
- 33:53wouldn't know what to do with it. If I
- 33:55put 5 million into ads, I wouldn't be
- 33:57able to handle the commercial follow-up
- 33:58, have the right trainers, or monitor
- 34:00the quality of the training. I think
- 34:02it's very hard, much harder than with a
- 34:04product where things scale easily. So,
- 34:05all things considered, we prefer to go
- 34:07a bit more step-by-step. Besides, our
- 34:09competitors are doing much the same. We
- 34:11haven't seen any training providers
- 34:12raise huge amounts of money like that;
- 34:14they all have this same problem of
- 34:15having to execute behind the scenes. So
- 34:17, it's completely tough. You mentioned
- 34:20several clients, notably Vinci, and
- 34:21Clarin as well, who are therefore not
- 34:23in real estate. At what point did you
- 34:26make the switch from focus to a gradual
- 34:28diversification of your client typology
- 34:30, including B2B and B2C? I'm asking you
- 34:33this because it's a real issue for many
- 34:35entrepreneurs: should I stay focused on
- 34:37a very specific niche, or allow myself
- 34:39to open up to a wider niche to
- 34:41potentially address a larger pool? It's
- 34:44a super interesting question, and I
- 34:45like to describe it this way. Actually,
- 34:47we stayed niche in real estate for 6
- 34:49months, I would say. Sure, you had
- 34:51clients here and there coming from
- 34:52elsewhere, but we focused on real
- 34:54estate for 6 to 8 months. The point of
- 34:57that, I think it's very important to do
- 34:58it that way, because when you focus on
- 35:00a niche, you become an expert in that
- 35:02niche, and all your prospects know
- 35:03you're an expert, so they sign with you
- 35:05faster. So, we were real estate experts
- 35:08: developers, social landlords, real
- 35:10estate agents, notaries. Those were our
- 35:12four main areas originally, so very
- 35:13quickly all the notaries want to work
- 35:15with you, all the social landlords want
- 35:16to work with you, and all the
- 35:17developers, and so on. So you grow much
- 35:19faster than our competitors who were
- 35:21generalists. And since you grow fast, I
- 35:23always use this example: we reached 35
- 35:25to 40 employees when our competitors
- 35:27were still at 6, 7, 8, or 9, and
- 35:28honestly, I don't know if we're better,
- 35:30I'm not sure. They are very good; I
- 35:33know a few who are very good. Do we
- 35:34have a better brand than them? I'm not
- 35:36sure either. Well, you can have a much
- 35:38better product than the other, so you
- 35:40grow faster. But we are good, I think,
- 35:42but I think they are very good too, so
- 35:44that's not what made us go from 35 to
- 35:4670. It's that we were very niched and
- 35:47said, "We won't diversify, we will
- 35:49execute what we do well and grow," but
- 35:51then you reach that moment where you're
- 35:5230 or 35, and the market starts to
- 35:54become a bit more mature. So then you
- 35:56start having larger companies
- 35:58interested in AI that ask themselves
- 35:59who they should have as a partner. And
- 36:01there are two choices on the market:
- 36:02companies with 7 people or companies
- 36:03with 35. And so, what did we get in our
- 36:06first clients? Agribusiness companies
- 36:07started coming to us and asking, "Well,
- 36:09don't you do agribusiness?" I would say
- 36:10, "Yes, we don't do agribusiness."
- 36:11However, we have 40 people at Misteria,
- 36:13while our competitors have 6. So if we
- 36:14work with you, you're a bit large, I
- 36:16have the capacity to provide you with
- 36:1730 days of training a month starting
- 36:18tomorrow. Our competitors can't do that
- 36:20. And so then they say, "Oh, yeah,
- 36:21that's not bad." So you say that to one
- 36:23agribusiness firm, then two, then three
- 36:25, and boom, you open a new vertical,
- 36:26and in fact, you grow that vertical
- 36:27quickly and soon you hit 60 employees
- 36:29while your competitors are at 12. And
- 36:32then you have Vinci or Clarins arrive
- 36:33and say, "Well, I need to train 600
- 36:35people all over the world at once, but
- 36:36I need a huge volume of trainers." I
- 36:38say, "Well, we have 60 people." So, in
- 36:40fact, your niche allows you to grow
- 36:41very quickly, to reach a sort of
- 36:43critical mass which then allows you to
- 36:44branch out. And for us, that's been the
- 36:47main strategy, which we're still doing
- 36:49in real estate, which is still 60%of
- 36:51our business.
- 36:53And we aren't going to let that go,
- 36:54because we like it, it's a huge market,
- 36:55and we're good at it, but also because
- 36:56it allows us to grow, grow, grow our
- 36:58company and become more structured. You
- 37:00see, we wouldn't have had a CFO or an
- 37:02HR director if we hadn't grown so fast.
- 37:04But having a CFO, an HR director, and
- 37:06so on means that new prospects tell us,
- 37:07"Oh yes, you are a serious company, you
- 37:09are structured." So even if I'm in
- 37:10aerospace and it's not your specialty,
- 37:12I know you'll assign someone to me who
- 37:14knows aerospace well and who will
- 37:16follow me from A to Z and so on. So
- 37:18that's a big piece of advice that I
- 37:19give. I would say be niched at the
- 37:21beginning, grow to be bigger than your
- 37:23competitors because you are niched, and
- 37:25then you have all the weapons to, well,
- 37:27I don't know, to go and attack other
- 37:29markets. But I think you can see it in
- 37:31plenty of other areas. It has to be
- 37:33like that, you know? You you build a
- 37:35team a bit on a on a specific vertical
- 37:37and in fact you reach a critical size
- 37:38that allows you to go elsewhere, you
- 37:40know. So that that was the strategy we
- 37:42used and that we are going to continue
- 37:43to use even if today, for that matter,
- 37:45we no longer present ourselves at all
- 37:46as a specialized real estate firm. We
- 37:48have specialized verticals. However,
- 37:49Mystoria is very generalist and we do
- 37:51everything there. And actually, that
- 37:53will allow me to ask you a question
- 37:54about the acquisition side. Because
- 37:56since it's very diversified, both in
- 37:59the range of offers and in the target
- 38:01panel, because you won't prospect or
- 38:03close a deal the same way with Vinci or
- 38:05Clarence as with an individual or a
- 38:07business owner, a fishmonger who wants
- 38:09to train in AI. How do you structure
- 38:13your offers? What exactly do you offer?
- 38:16Do you offer it directly, just
- 38:18front-end on the site, and you go, you
- 38:19choose, you click, and you start an
- 38:21adventure? Or does it happen
- 38:23differently?
- 38:24So first of all, our company structure
- 38:25is in three parts. It's B2B consulting,
- 38:27B2B training, B2C training. Those are
- 38:30really the three products of Mysteria.
- 38:31And how do we sell it? A lot of
- 38:33webinars, a lot a lot of webinars. I do
- 38:35webinars every week but with zero paid
- 38:37acquisition. Actually, our newsletters
- 38:39—yes, newsletters, because we have
- 38:41Mysteria, we have the one for real
- 38:42estate agents, and we have the one from
- 38:44my partner's company, which by the way
- 38:45has now been acquired by Mysteria,
- 38:47which also reaches 30,000 people in the
- 38:49real estate sector. So, we have our
- 38:51three newsletters that draw about
- 38:53100,000 people that we invite to
- 38:54webinars very regularly, and in fact,
- 38:56you realize that people send their
- 38:57friends, or sometimes we do little
- 38:59promos, we tell them: "Well, if you
- 39:00invite three people who don't know
- 39:02Mysteria, at the end we give something
- 39:03away," etc. So, in fact, you pull in
- 39:05lots of people who don't even read the
- 39:07newsletter. So we do a lot of free
- 39:09webinars, we present our strategy, and
- 39:11we're going to keep that up—we give
- 39:13you free info. So we do an hour of
- 39:15content on AI, and at the end we tell
- 39:17you if you want to be accompanied, you
- 39:18can ask us, and we'll be happy to
- 39:20accompany you. That's the big growth
- 39:21strategy. First, second growth strategy
- 39:23is to be very niche by profession. Even
- 39:26now. Today, we have 6 or 7 professions
- 39:28where we are still very strong. And
- 39:29niche by trade means, for us, attending
- 39:31the trade shows that matter. Misteria,
- 39:32for example, doesn't go to AI trade
- 39:34shows. All my competitors do AI trade
- 39:36shows. I, however, go to the public
- 39:37housing trade show. At the public
- 39:38housing show, everyone says, "But what
- 39:39are you doing here?" Well, actually,
- 39:41there are 600 to 1,000 public housing
- 39:43firms in France with 100 to 1,000
- 39:44employees that need AI just like
- 39:46everyone else. So when I go there, I am
- 39:48the AI expert for the public housing
- 39:50sector. Same for notaries: we attend
- 39:52the Notaries' show, we do real estate
- 39:54developers 'shows, we do the Mayors'
- 39:55show for local authorities, but we
- 39:57don't do Big Data shows, we don't do
- 39:58Vivatech—though we should, and we
- 40:00probably will one day. But given the
- 40:01choice today, we invest much more in
- 40:03these other shows. So we see a lot of
- 40:06growth from acquisition happening this
- 40:08way. Trade shows, webinars, word of
- 40:11mouth—it’s nearly half of our
- 40:12acquisition. One out of every two
- 40:14people says, "My friend works with you,
- 40:16he told me to reach out, so I'm working
- 40:17with you."
- 40:18Because the ecosystem is still small,
- 40:20we have that advantage. It feels like
- 40:22everyone is doing AI, but in reality,
- 40:24companies are just starting out, so
- 40:25naturally, whenever they have questions
- 40:26, they look around and ask, "Do you
- 40:28know someone who helped you?" And
- 40:29Misteria comes up a lot. SEO as well.
- 40:32That's starting to pick up; we're
- 40:34beginning to rank well. Those are our
- 40:36main strategies. But the acquisition
- 40:38cost is virtually zero. We launched our
- 40:40first Google ad at the beginning of
- 40:41September. It's been 9 days of running
- 40:44Google ads for small amounts.
- 40:46We also made our first paid acquisition
- 40:48for a webinar this month. So before
- 40:50that, we had zero paid acquisition.
- 40:52I’m speaking primarily about B2B here
- 40:53. In B2C, we started doing paid
- 40:56acquisition a little earlier, but it's
- 40:58still quite small for us. Those are the
- 41:00main channels. And well, LinkedIn—I
- 41:02have about 27,000 followers, I think;
- 41:04I've been posting every day for 2 years
- 41:06, and it still drives a lot of traffic.
- 41:09So, you see, we have a real organic
- 41:11acquisition strategy for now. But trade
- 41:13shows, you still have to pay for them,
- 41:14they aren't free, even if the amounts
- 41:15are small. So, trade shows plus organic
- 41:17, that's the bread and butter for us
- 41:19today. And often, and here I'm talking
- 41:21more B2B than B2C, the person arrives,
- 41:23"I want training," well, that's perfect
- 41:25, we're CPF eligible, so we train a lot
- 41:27of them using CPF. We're hitting some
- 41:28pretty crazy numbers right now. Which
- 41:30is great, the CPF allows for that.
- 41:31We're pushing the government to steer
- 41:33CPF funds more toward AI because it's
- 41:35really important for the country. So
- 41:37B2C is pretty standard. They watch the
- 41:39webinar, they ask, they use their CPF,
- 41:40and they train with us. On the B2B side
- 41:42, the request often comes in saying, "I
- 41:44have no idea where to start with this
- 41:46AI topic." Companies tell me that. They
- 41:48see us in a webinar, at a trade show,
- 41:49or where I mentioned, and they ask, "
- 41:51Where do I start to get a little bit
- 41:53going?" So, what we push very often is
- 41:55a quick audit to start. We say, "We'll
- 41:58audit you, see what you need, create a
- 41:59roadmap, and then roll out training,
- 42:00some custom development, and so on."
- 42:02But today, that's mainly it. You have
- 42:05very few companies that come to you
- 42:06saying, "Okay, I have an AI training
- 42:07plan in mind, I'm looking for a
- 42:08provider to do this, that, and the
- 42:09other." That's quite rare. Most are
- 42:11like, "Guys, we need help, we don't
- 42:12know how to approach this thing."
- 42:14Yeah, we'll get to that. But before
- 42:16that, I want to jump on a point you
- 42:17raised that's super interesting, which
- 42:19is the trade show aspect. Uh, it's true
- 42:21you might think, "Well, Vivatech, the
- 42:23AI show, etcetera," and so on. However,
- 42:26this strategy is better than being a
- 42:28shrimp in the ocean; you become a whale
- 42:30in a pond. It's a bit like blue ocean,
- 42:32red ocean. Uh, you show up there,
- 42:34you're the only AI player, talking
- 42:35about the HLM trade show or other shows
- 42:36you mentioned. It's pretty crazy, and I
- 42:39think many should take inspiration from
- 42:40this type of strategy if, in fact,
- 42:42you're a service provider who can help
- 42:43one or more business sectors rather
- 42:45than just going to your own trade show.
- 42:47Clearly. Clearly. And people really
- 42:49like that too, you know. Me, whether
- 42:50it's at trade shows or even outside of
- 42:52trade shows, I spend all my time out in
- 42:53the field. Actually, I'm not in the
- 42:55office enough; that's what I blame
- 42:56myself for. I would like to stay there
- 42:58more. You see, yesterday I was in Niort
- 43:00, tomorrow I'm in Toulouse. The day
- 43:02after tomorrow, I'm in Toulouse. I'm
- 43:03moving around all the time. Not to have
- 43:05fun, but to see clients. Actually, they
- 43:06really appreciate it and I find that
- 43:07normal. They are tired of the little
- 43:09Parisian who explains life from his
- 43:11Paris office, telling people what to do
- 43:13. No, I tell them I'm not going to do
- 43:14that, I'm coming to see you in person
- 43:16and we'll talk. And actually, when you
- 43:17do that, people are so much nicer, they
- 43:18put you in touch with a lot of people.
- 43:20That is our real strategy. We really
- 43:22try to avoid the thing where I'm a bit
- 43:24of an AI star in Paris, which I'm not,
- 43:26but I could very well act like it, you
- 43:28see. And no, we try to avoid that
- 43:29absolutely. We go into the field, we go
- 43:31to trade shows, we meet people, we get
- 43:33to know them well, and we don't
- 43:34hesitate, however, to tell them: "Look,
- 43:35things are going very well between us."
- 43:37You're going to say: "I'm, well, I tell
- 43:38the person, hey, if you can put me in
- 43:39touch with one of your colleagues, that
- 43:41would be great, we're looking to expand
- 43:42here a bit," and everyone does it. When
- 43:43things are going well with a client and
- 43:45you are nice, everyone accepts you. And
- 43:47the conversion rate is 100%. With a
- 43:50personal connection, the conversion
- 43:51rate really is 100%. So we rely on that
- 43:53a lot and I think that, indeed, that
- 43:54might be a mistake of the French Tech.
- 43:56I saw recently, I think it's Payfit or
- 43:58Qonto, I don't remember, maybe Payfit,
- 44:00the CEO who said the company is almost
- 44:0210 years old, created in 2015-2016 or
- 44:04is about to be 10. And the CEO says, a
- 44:06year ago, "Okay, let's go, for the
- 44:08first time, I'm going to do a tour of
- 44:09France to visit my clients." And I said
- 44:11to myself, but what have you been doing
- 44:12for 9 years? You see, obviously, go see
- 44:14the clients. Now, it's easier to say
- 44:16that from my sofa than anything else,
- 44:17right, I'm not criticizing either. But
- 44:18it's quite interesting and, besides, I
- 44:20think in his place, I would have done
- 44:21the same. I didn't really go see my
- 44:23clients for 10 years because the thing
- 44:25was, we all raise a lot of money, we
- 44:26spend heavily on ads, and we close.
- 44:29It's fine, but I feel like there's a
- 44:30bit of a shift today where people want
- 44:32to see you, and we have plenty of
- 44:33competitors. So if companies don't see
- 44:35me, I'm no better at ads, I'm no better
- 44:37at whatever else. So they need to see
- 44:39me so they can trust me. One, it's
- 44:41better for selling, and two, for
- 44:43improving your services. When you're
- 44:45selling services. I tell them, I ask
- 44:46them concretely, what do your employees
- 44:47actually do? You know, what's their
- 44:49daily problem? I go into the teams, I
- 44:51go to their offices very often, and I
- 44:53actually dig up gold mines for my
- 44:54consultants when I bring them tons of
- 44:56info. And so even they have that rule.
- 44:58I tell them, when a client wants to see
- 45:00you in Brest, you go to Brest. I don't
- 45:02care if we lose time, we go see them in
- 45:03Brest. And you have a trade show, a
- 45:06notary fair, you know, it's not the
- 45:07sexy thing in Paris or whatever. It's
- 45:10sometimes in lesser-known cities,
- 45:11sometimes it's not very big. Well, we
- 45:13go, we don't care. We see our clients,
- 45:14they're happy to see us, we're happy to
- 45:16see them, and I find it's a sales
- 45:17strategy that also works very, very
- 45:18well.
- 45:19And it's one that's sincere, you see.
- 45:21It's not just ADS anymore, it's the ADS
- 45:23sales strategy. You might tell yourself
- 45:25you're trying to reach people's
- 45:26emotions and so on, and that's the game
- 45:27. Why not? But here, when you go to see
- 45:29them, it's sincere, you talk properly
- 45:31with them, you say, "Look, I can do
- 45:32this for you," does that work or not?
- 45:33We make a deal. It's more of a, let's
- 45:35say, grassroots business approach. I
- 45:37like it completely. You also say that
- 45:40you're in a very competitive market.
- 45:42Yeah, which is true. Today, anyone,
- 45:44like you said, could just sit on their
- 45:47couch and criticize. But ultimately,
- 45:49today, anyone on their couch, in their
- 45:51room, or anywhere else with a computer
- 45:52and an internet connection—wherever
- 45:54they are in the world, regardless of
- 45:56their age or background—can create a
- 45:58website and start an AI agency with AI
- 45:59agents or just a few people. So indeed,
- 46:01there are an enormous number of players
- 46:02in the market. You also mentioned
- 46:04LinkedIn, where there are tons of
- 46:06experts or pseudo-experts present. The
- 46:08idea isn't to criticize or point
- 46:10fingers, but rather: beyond this
- 46:12strategy you mentioned, which is
- 46:14hyper-interesting—the ground-level,
- 46:16down-to-earth side, the contact with
- 46:19the client—how do you differentiate
- 46:21yourself in an extremely competitive
- 46:23market with low barriers to entry,
- 46:25relatively low margins, and
- 46:27extraordinary growth?
- 46:29It’s tough, it’s tough. So, the big
- 46:31pillars: if I had to start over in
- 46:33another ecosystem, I’d tell myself:
- 46:35you need to see people because in
- 46:37consulting, they hire you for a project
- 46:39because they know you, not just because
- 46:41you're the best.
- 46:42So you have to see people,
- 46:44that’s the first thing. And you have
- 46:47a scale issue in agencies, regardless.
- 46:50Agencies with two, three, or even ten
- 46:53people can handle B2B for small
- 46:55companies. But when you’re dealing
- 46:57with a company of a thousand people, I
- 46:58don't know if they're right or not, but
- 46:59I think so. They want to have a very
- 47:01structured company in front of them, a
- 47:02very structured provider. So, there is
- 47:04still a race to scale, which we are
- 47:05currently winning against some
- 47:07competitors. So that really helps us.
- 47:10Our very pragmatic approach—I'll
- 47:13speak more about AI—is very popular.
- 47:16I have plenty of competitors saying, "
- 47:18We’re going to create agents for you,
- 47:19we’ll automate 70%of your business."
- 47:21I tell them, we’ll set up ChatGPT,
- 47:23and you’ll all save an hour a day.
- 47:24Even though we obviously do more than
- 47:25just that. That’s more of the promise
- 47:27, and it’s much more down-to-earth.
- 47:29And I tell them if we start now, in a
- 47:30month, your 1,000 employees will be
- 47:31trained. Now, that’s starting to get
- 47:32a bit concrete. So, we try to be
- 47:34concrete and maintain a pragmatic
- 47:35approach. That’s more in relation to
- 47:38our big competitors, because compared
- 47:40to the small ones, I play the size and
- 47:42support card. I say, look, if you train
- 47:43with a freelancer, they won't follow
- 47:45you for the next two years. For us,
- 47:47you’re a 100-person company. You're
- 47:49hesitating between us, with our 100
- 47:50employees, and a 10-person firm. Well,
- 47:53since we're 100, we'll handle your
- 47:54training, your consulting, and give you
- 47:55a packaged offer. And since people find
- 47:57it hard to imagine the business needs
- 47:59of tomorrow regarding AI, we tell them
- 48:01we'll be on the front line no matter
- 48:03what to address them. If tomorrow we
- 48:06feel you actually need to do some
- 48:07AI-driven SEO, or need some advanced
- 48:09training or a bit of consulting here
- 48:10and there, we know you so well and have
- 48:12already supported you on so many things
- 48:14that we'll be able to support you on
- 48:15that too, because we're big. So against
- 48:18smaller competitors, so to speak, I
- 48:19play that card, and against the very
- 48:21big ones like OnePoint, Artefact, and
- 48:22the like, we also try to hold our own,
- 48:24and there we play the pragmatism card.
- 48:26We say, actually, that BCG strategic
- 48:27mission at € 500,000 to tell you that
- 48:29AI will help you write meeting minutes
- 48:31...At Nouvel, it costs € 20,000 and
- 48:33we provide something very pragmatic.
- 48:35That was really the focus from the
- 48:37start, deciding we would do practical,
- 48:39hands-on missions. I've seen many
- 48:41missions from big consulting firms pass
- 48:43through our clients 'offices. First off
- 48:45, they sell it for a fortune because
- 48:46they don't get out of bed for €
- 48:4710,000. So they sell it for a fortune
- 48:49and it's theoretical: there will be
- 48:51productivity gains here, there will be
- 48:52a reorganization of the company there,
- 48:54and so on. And my clients tell me, "I
- 48:56received this from X consulting firm,
- 48:58what am I supposed to do with it?"
- 49:00Whereas I, at the end of the mission,
- 49:01tell them, "Look, we've audited your
- 49:03development team." For your development
- 49:04team, we found 30 use cases. For each
- 49:06use case, we made a demo video of how
- 49:08they could implement it. We identified
- 49:09the tool or tools that could be used.
- 49:11How much it would cost you in licenses
- 49:12if you took one, if you took five. By
- 49:14the way, we also told you how many
- 49:15licenses you actually need. And there,
- 49:17actually, at the end of it, the company
- 49:18has something super concrete. So that's
- 49:20the card we play against the big
- 49:21players. It's when we're dealing with
- 49:23big corporate clients, where it's
- 49:24either us, or a OnePoint, or an
- 49:26Artefact, or an EY facing them, that we
- 49:27tell them: "Well, with us, you won't
- 49:29get the Artefact or OnePoint stamp,
- 49:30you'll get a MisterIA stamp that's
- 49:32worth much less on the market, but
- 49:33you'll get something extremely concrete
- 49:34." So that still appeals to many, to
- 49:37many big companies. That's how we try
- 49:39to differentiate ourselves a bit
- 49:40between the two, between the very small
- 49:42freelancer training companies and the
- 49:43very large firms doing massive
- 49:44development plans.
- 49:45Very clear. Earlier, you mentioned that
- 49:48many companies don't really understand
- 49:50how to implement AI in their workplace,
- 49:53and you said, "Well, click here, get an
- 49:55audit, and we'll explain it to you."
- 49:58Let's imagine that the company watching
- 50:00this episode doesn't have the immediate
- 50:02capacity to call us or get an audit.
- 50:05What would be the first major general
- 50:07pieces of advice you'd share with any
- 50:09company, say with a turnover between
- 50:11500k and 100 million euros, that would
- 50:13like to implement AI in their daily
- 50:15routine? Well, the first reflex for
- 50:18those who don't necessarily have the
- 50:20budget for consulting is to go ahead
- 50:21and get ChatGPT licenses for your
- 50:22employees. Be transparent about AI
- 50:25usage, meaning authorize your employees
- 50:27to use it, but understand what they're
- 50:29doing with it. I think that's quite
- 50:31important. There are still issues with
- 50:33data, etc. So, be careful. So, get them
- 50:35licenses, and take a look at what
- 50:36they're doing. Set up a small AI
- 50:38committee—many are already doing it
- 50:40—with 4 or 5 people in the company
- 50:42who are interested in the topic, and
- 50:43meet once a month for a morning to
- 50:45reflect on what can be implemented, and
- 50:47so on. You don't need to do big things,
- 50:49but at least small use cases will
- 50:50emerge. That works well. Large
- 50:52corporations do this a lot. For example
- 50:54, we have an AI ambassador support
- 50:55service; many large groups tell us, "
- 50:57I've created 50 AI ambassadors, and
- 50:58that's their role, to do what I just
- 51:00described." However, they are supported
- 51:01by MisterIA because the stakes are high
- 51:03, and so on. But that works well. So,
- 51:05set up a small AI committee. After that
- 51:07, consume content on YouTube, but
- 51:09there's so much content, it's hard.
- 51:11Well, it's the classic advice, but it's
- 51:14hard to really know what to do just by
- 51:16consuming content on YouTube. Um,
- 51:20otherwise, you have coaching programs
- 51:23that we know well, which work quite
- 51:24effectively. I have quite a few CEOs
- 51:27who tell me they don't want to spend 10
- 51:29, 20, or 30k on an audit. However, they
- 51:31do want to be guided for 3 or 4 months
- 51:33to really understand the subject. So we
- 51:35do a lot of coaching like that, where
- 51:36you're one-on-one with your coach and
- 51:38you just get things moving. So these
- 51:40are things that work well. So, overall,
- 51:43that would be the advice I'd give to
- 51:44people who don't have too much budget:
- 51:46start with a short training day, which
- 51:48isn't very expensive and already allows
- 51:49you to cover a lot of ground. The idea
- 51:52of a committee is really interesting.
- 51:53It effectively allows you to empower
- 51:55one, two, or three people in your
- 51:57company and tell them, "Stay informed
- 51:59on the subject once a month." You meet,
- 52:01and based on the daily use cases we
- 52:03have, we see how we can gradually
- 52:04implement it. In my experience, it
- 52:06might seem a bit abstract to some who
- 52:09haven't adopted AI yet. As crazy as it
- 52:11might seem to you or me, since we are
- 52:13immersed in AI every day and we talk
- 52:15about it a lot on entrepreneurance.com.
- 52:18Can you explain these famous concrete
- 52:20use cases where you say, "You're going
- 52:22to save an hour a week per employee,
- 52:24maybe even an hour a day per employee
- 52:26at certain times"? Sometimes it's 15%,
- 52:3020%, 30%more productivity, not to
- 52:32mention the costs you save because,
- 52:34ultimately, ChatGPT is only 20 dollars
- 52:37a month, which is nothing, but you pair
- 52:39it with other tools. You mentioned
- 52:43Jasper, just to name one, but there are
- 52:45so many others that we can use. We'll
- 52:47talk later about all the tools and how
- 52:49not to fall into the frenzy of using
- 52:511,500 of them to no avail. But
- 52:53basically, in concrete terms, how do we
- 52:55manage to save even just one hour a
- 52:57week for each employee?
- 52:58Yeah. Well, in terms of use cases that
- 53:01work well, for the entire sales side of
- 53:02companies, there's a lot to be done.
- 53:05All our sales proposals are done in one
- 53:07click and they are personalized, they
- 53:09are interactive, they are beautiful,
- 53:11they are detailed, and that is
- 53:12impossible to do without AI. So, a bit
- 53:15like I was telling you, I take my
- 53:16outreach, I send it to my to my Claude
- 53:17—I use Claude a lot—which creates a
- 53:19sales proposal for me in the form of a
- 53:20website. You send that to the client,
- 53:22nobody else is sending them that. There
- 53:24is a huge "wow" factor, and they have
- 53:25all the information. So, just with that
- 53:27, you’ve saved a lot of time. Meeting
- 53:29minutes. So, I was talking about video
- 53:30calls, but even in-person meetings with
- 53:31a prospect. Me, as soon as I finish a
- 53:33physical meeting, I leave, take out my
- 53:35phone, and I dictate a summary. And for
- 53:37the summary, I tell it, "Here is
- 53:38everything that happened, write a recap
- 53:39email and a WhatsApp recap for my
- 53:41partner."
- 53:41It's simple, but it's done in a minute
- 53:43and a lot of information gets shared.
- 53:45So that saves a lot of time in the long
- 53:46run. Then, as for all your analyses, it
- 53:49really depends on the profession, you
- 53:51know. The use cases are all very
- 53:53job-specific, but for all your document
- 53:55analysis, any document I want to
- 53:57analyze, I send it to an AI—to Claude
- 53:59or ChatGPT—and I tell it, "Summarize
- 54:00this for me, analyze this, compare
- 54:02these two docs." You see, if you have
- 54:04if you work with contractors, send five
- 54:06quotes from contractors to Claude and
- 54:08say, "Make me a summary table." Those
- 54:10are things that take up a lot of time.
- 54:12Uh, sometimes all the email management,
- 54:13for me 100%of emails are answered with
- 54:15AI, 100%. I have a ChatGPT shortcut, I
- 54:18talk to it, and actually I see the
- 54:19email, it sees it too. I I click
- 54:21Option-Space, and it brings up a tiny
- 54:24ChatGPT window. I click the microphone,
- 54:26I say, "Okay, reply to Alex to tell him
- 54:28I'll be there tomorrow, that I'll
- 54:29arrive at 11:00 AM, and that and that I
- 54:31want to have some tea." And in fact, it
- 54:33writes the email right before my eyes,
- 54:35I copy, I paste, I send. It saves me a
- 54:36crazy amount of time. Just that alone
- 54:37is definitely 40 minutes a day. So you
- 54:40have a few cases like that, email,
- 54:41sales teams, well finance teams, you
- 54:43have a bit fewer actually, you see, I
- 54:45have fewer use cases in mind for the
- 54:47finance side. Uh HR teams, every
- 54:50company has HR teams, preparing your
- 54:51interviews, writing your job
- 54:53descriptions, comparing a resume with a
- 54:55job description. I send both and I say,
- 54:57I send the job description and I say,
- 54:59out of these five resumes, which ones
- 55:00best match the job description and why?
- 55:02Uh it does it for us all the time, you
- 55:03see. It allows us to have a single HR
- 55:05manager who does daily interviews,
- 55:07which she could never do otherwise. But
- 55:09actually, she uses AI that helps her do
- 55:11all of that. She, for example, gives
- 55:12the job description and resume to the
- 55:13AI, and tells it, "Prepare the 15
- 55:15interview questions I need to ask."
- 55:16H
- 55:17It’s unbeatable. You have plenty of
- 55:19little use cases like that. And in all
- 55:20of that, even just picking one or two
- 55:22and doing them well, you’ll save an
- 55:23hour a day. 100%per day. That’s
- 55:25really the
- 55:26equivalent. And come on, let’s
- 55:28include the whole communication side
- 55:29which works very well. And
- 55:30communication is important. Careful, "
- 55:33I’ll write my LinkedIn post with AI,"
- 55:35it’s good, it can work, but you
- 55:36always have to go further in the use
- 55:38case. I often say, for example,
- 55:40regarding the LinkedIn post case,
- 55:42it’s good to write it with AI, but
- 55:44instead tell your ChatGPT: "I want to
- 55:45write this LinkedIn post, I have this
- 55:47idea in mind, help me turn it into a
- 55:49series of 10 LinkedIn posts.""I want my
- 55:5110 LinkedIn posts to have a common
- 55:53thread.""I want three of my 10 LinkedIn
- 55:55posts to be aimed at my prospects,
- 55:56three to attract the curiosity of my
- 55:58clients, and three to attract the
- 56:00curiosity of my future employees for
- 56:02employer branding.""And in these 10
- 56:04LinkedIn posts, I want you to give me
- 56:05little examples each time, and in each
- 56:07post, tease the one for the next day a
- 56:08little bit." You see, you can play with
- 56:10it like that and you’ll get it to
- 56:12create something much more structured,
- 56:13much more powerful, rather than just
- 56:15telling it: "I have a post idea, write
- 56:17it for me." It works. I write
- 56:18everything with it, so it works, but we
- 56:20can take it further. You see, today
- 56:22these use cases are somewhat well-known
- 56:23. However, I don't see anyone using it
- 56:24the way I do, like, "Okay, I want 10 of
- 56:26these." I want each LinkedIn post to
- 56:28link to a Twitter post I'll make the
- 56:30same day as my newsletter; you have it,
- 56:32honestly, we have a newsletter. I
- 56:34always tell him, "Take my newsletter,
- 56:36break it into a bunch of posts,
- 56:37announce something, dig a little deeper
- 56:39with it," and in terms of communication
- 56:41, you'll save a lot of time, too. Why
- 56:43is it that many entrepreneurs lack the
- 56:45desire to dig deeper into the subject
- 56:47or push things further? Because it
- 56:49seems obvious when you explain it, just
- 56:51like saying, "Well, first ask one
- 56:53question at a time, and then dig deeper
- 56:55, develop it," and. And it's true that
- 56:58you have much more advanced tools,
- 57:01while still staying on ChatGPT, just
- 57:03within its "deep search" features or
- 57:05others, where I've already had the case
- 57:07: I push it so much to the limit that
- 57:10it takes me several days to get the
- 57:12full summary out. And I have to follow
- 57:15up every 2 hours so I can say, "So, are
- 57:15you done thinking?" And boom, and boom,
- 57:16and boom, it spits out a document up to
- 57:1780 pages long. That's how I built the
- 57:19vision for entreprenance.com for the
- 57:21next 5 years. For example, 80 pages
- 57:22that were written between ChatGPT and
- 57:24me, where it challenged me, and then I
- 57:26just submitted it to different people
- 57:28around me to see if it was compelling.
- 57:30But that would have never been possible
- 57:31. It would have taken me 4 months to
- 57:33write that document. Completely,
- 57:34absolutely. Why don't people dig deeper
- 57:36? Is it a lack of time? I think that's
- 57:38still the problem. Wait, service firms
- 57:40exist largely because people don't have
- 57:41the time. It's always the same; you're
- 57:43buying two things from them. Either you
- 57:45buy their expertise, or you buy the
- 57:46time of junior consultants that you
- 57:47don't have yourself. I think that's a
- 57:49bit true in every business, and so we
- 57:51obviously bring the expertise a bit
- 57:52because not everyone knows all the
- 57:53tools, and so on. And also, I have
- 57:55plenty of entrepreneurs who tell me if
- 57:57I don't have someone with me, if I
- 57:59don't block out a day-long slot and
- 58:01have someone with me to explain it,
- 58:03I'll never do it. And I understand them
- 58:06, and I’m the first to admit it;
- 58:07sometimes I get coached on things where
- 58:08I could tell myself, "Look, I have
- 58:10enough perspective to go on YouTube,
- 58:11sit down for an hour, look up a bit of
- 58:12GPT, a bit of all that," and learn, but
- 58:13I’ll never do it. So I don't think
- 58:16it's out of ignorance, it’s just that
- 58:17they don't have the time. So you have
- 58:19to find the time, compel yourself,
- 58:20force yourself. Uh, that's the first
- 58:22thing. Secondly, there’s a big bias
- 58:24with AI: you save time very quickly.
- 58:27You save a little time very quickly,
- 58:28and so you often stop there. You do
- 58:30your research on ChatGPT, you write a
- 58:32few emails, and you’re already saving
- 58:34a lot of time. So you often tell
- 58:35yourself, "Well, it's good, I already
- 58:36have a good handle on it." And you
- 58:38never make the effort to say, "Oh wait,
- 58:39actually, maybe I could save not just
- 58:41one hour, but 3 or 4 hours." That is a
- 58:42big, big problem we see in companies,
- 58:44for example, where I have people tell
- 58:45me, "No, it's fine, my employees all
- 58:47use ChatGPT, it's great." Or others who
- 58:49...I insist a lot on saying, "GPT saves
- 58:51you time, yes, but it also provides a
- 58:53lot of quality." And it's really hard
- 58:55to make employees understand that it
- 58:57will improve quality, because the time
- 58:59savings are just too visible, and it's
- 59:01too easy to save time. So that’s also
- 59:02a bias I think is somewhat important,
- 59:04which is telling yourself, "Oh yeah,
- 59:05okay, but it does other things too."
- 59:07Kind of like, I don't know, the Office
- 59:08suite; you see, we could say that with
- 59:10Outlook, we all just write our emails
- 59:12on it. I’m quite certain that we
- 59:14could push productivity much further in
- 59:15Outlook. Actually, it’s so practical
- 59:17for writing emails, and we don't have
- 59:18the time to look into it too much. So
- 59:20we only use that part of the Office
- 59:21suite. Yet, with AI, we could go much
- 59:23further. So that’s, that’s a bit of
- 59:25a hurdle. And by the way, that was the
- 59:27big aspect of ChatGPT. For two years or
- 59:29a year and a half, you couldn't do
- 59:31internet searches on ChatGPT. I used
- 59:33Perplexity a lot, which still works
- 59:34very well.
- 59:36And I'm quite convinced that the fact
- 59:38that ChatGPT can search the internet is
- 59:40a basic feature for them. It's easy to
- 59:43develop. Why didn't they develop it for
- 59:45a year and a half? It’s not because
- 59:46OpenAI isn't good or because they lack
- 59:48engineers; they have everything they
- 59:49need. I think they didn't want us to
- 59:51fall into the trap of turning it into
- 59:52another Google. I think they said: "
- 59:55They're right, we've created a tool so
- 59:57powerful that we can take so far that,
- 59:58if from day one, we let Google search,
- 1:00:00Internet search happen on it, everyone
- 1:00:02will use it like that and stop and
- 1:00:03never dig any deeper." And so for two
- 1:00:06years, in my opinion, they refused to
- 1:00:08add it. And today, now that they’ve
- 1:00:09added it, well, we feel it’s one of
- 1:00:11the most used features, even though,
- 1:00:12heck, sure, you save a bit of time on
- 1:00:14research, but that's not the goal of
- 1:00:15ChatGPT. So that’s a bit of a bias.
- 1:00:17It’s that the first productivity gain
- 1:00:19is so fast, people stop a bit there
- 1:00:21without asking more questions, so we
- 1:00:23forget all the quality it brings. And
- 1:00:25you mentioned the fact of having an HR
- 1:00:28manager who actually uses AI on a daily
- 1:00:30basis to be able to bring quality and
- 1:00:32productivity to all these tasks. Yet
- 1:00:35you are hiring a huge number of people.
- 1:00:37Uh, we mentioned it earlier, now you're
- 1:00:40over 100 people. Uh.
- 1:00:43How many people do you hire on average
- 1:00:45per week? Why am I asking you this
- 1:00:46question? Admittedly, because I imagine
- 1:00:50the figure is interesting due to the
- 1:00:51growth, but also to make those
- 1:00:53listening to us realize how much AI can
- 1:00:55allow one to be much more efficient,
- 1:00:57much more effective, much more
- 1:00:58productive, and on top of that, with
- 1:01:00quality, because ultimately every hire,
- 1:01:02even just for external contributors,
- 1:01:04the freelancers you bring on, well,
- 1:01:06it's capital, because when you're in
- 1:01:08the service industry, it's your
- 1:01:09reputation you're putting in the hands
- 1:01:11of a freelancer.
- 1:01:13Every time for every mission, for every
- 1:01:15client.
- 1:01:15Completely. So we hire, it's not
- 1:01:17regular, we hire more by the quarter, I
- 1:01:20would say. It's not every week, you see
- 1:01:21, it's more by the quarter. Now, we're
- 1:01:23going to hire 15 people per quarter, 15
- 1:01:25to 20 people per quarter, I think,
- 1:01:26roughly.
- 1:01:27It's about that. And indeed, we only
- 1:01:29have one HR director, but that's only 5
- 1:01:31%of her time because she does all the
- 1:01:32staffing, prepares the training, and is
- 1:01:34also the head of pedagogy. So it's a
- 1:01:35lot, and I find that's the most
- 1:01:37striking part; where you're right is
- 1:01:38that service companies will be
- 1:01:40transformed by AI. And you have two
- 1:01:41things. You have the fact that we'll be
- 1:01:44able to do more, you see. Thanks to AI,
- 1:01:46as a service company, back in the day,
- 1:01:48I might have only been a training
- 1:01:49company. Today, I can do some
- 1:01:51consulting, I can do some coaching and
- 1:01:53support because we are very
- 1:01:54well-equipped, and secondly, it's going
- 1:01:56to blow up the margins. That's going to
- 1:01:58be the big topic, though. There's a
- 1:01:59margin issue in the service sector
- 1:02:01that's coming, realizing that, well, we
- 1:02:03see it, we price ourselves twice as
- 1:02:05cheap as EY on equivalent assignments.
- 1:02:07For instance, we had a case on a
- 1:02:09digital audit assignment for a company
- 1:02:10where we were two to three times
- 1:02:12cheaper than EY, and the result, the
- 1:02:13mission is finished. The result is the
- 1:02:15same. And the big issue, well, the
- 1:02:17result, or at least the quality of the
- 1:02:19output, is the same. And the big issue
- 1:02:21is that all these firms are built with
- 1:02:23layers of consultants who are paid to
- 1:02:25make slides, paid to write meeting
- 1:02:27minutes, paid to interview employees,
- 1:02:29and so on. And that is over. And for us
- 1:02:31, for example, by using AI, it means
- 1:02:33that you have two choices. Either you
- 1:02:34skyrocket your margins, which we could
- 1:02:36have done, or, as we did, we lower our
- 1:02:37prices and skyrocket our volume. That's
- 1:02:39the choice we decided to take. And that
- 1:02:41is a major, major issue for service
- 1:02:42companies. If you are well-equipped
- 1:02:43with AI, I'll take a classic consulting
- 1:02:45assignment as an example. Our interview
- 1:02:46grids are made by AI; I say who to
- 1:02:48interview, what the topic is, and so on
- 1:02:50. Interview grid made by AI, we spend
- 1:02:52zero time. When we conduct an interview
- 1:02:54, we record it and then send that plus
- 1:02:56the grid, plus our analysis framework,
- 1:02:58and it gives us a full analysis of the
- 1:03:00person's interview. That's something a
- 1:03:02consultant normally does in a classic
- 1:03:03consulting firm. As for the slides, I
- 1:03:05don't know, maybe we price the cost of
- 1:03:07creating slides at about 5%of a
- 1:03:09mission's cost, whereas it's 30%.
- 1:03:11Usually, in my opinion, it's
- 1:03:12consultants who make all the slides.
- 1:03:14Well, for us, no, it's AI that makes
- 1:03:15the slides. We give it all the content,
- 1:03:17it makes all the slides, and it's like
- 1:03:18that the whole time. You have so, so,
- 1:03:20so many little things done with AI,
- 1:03:21which means that me, my partners—I
- 1:03:23have partners—they manage to handle
- 1:03:2510 projects at € 15,000 per project
- 1:03:27at the same time, you see, or even a
- 1:03:28bit more. And that, normally, you can't
- 1:03:30do that either in a traditional setup.
- 1:03:32So that is a huge, huge, huge topic.
- 1:03:35You have many who say that thanks to AI
- 1:03:36, people won't call on consulting or
- 1:03:38training firms anymore. I don't believe
- 1:03:39it. You will always need an expert; you
- 1:03:41always need to have an intermediary. In
- 1:03:43my opinion, we've been told for a long
- 1:03:45time that intermediaries would die out
- 1:03:46with the internet, but now there's AI,
- 1:03:47and I don't really believe it. On the
- 1:03:49other hand, clients are going to be
- 1:03:50much more demanding, that's for sure.
- 1:03:51And they'll say, "Why are you selling
- 1:03:52me this project for € 20,000 when I
- 1:03:53know you're going to do a lot of it
- 1:03:55with AI?" Well, we say: "Yes, we won't
- 1:03:56sell it to you for 20, we'll sell it to
- 1:03:57you for 10," and, indeed, we'll be very
- 1:03:59well-equipped behind the scenes. So
- 1:04:00that is a big, big, big productivity
- 1:04:02topic. All the companies doing SEO, the
- 1:04:04SEO agencies, there's an AI topic
- 1:04:06that's going to be huge. You won't be
- 1:04:08able to charge for an SEO management
- 1:04:10service today like you charged
- 1:04:11yesterday, because—I won't say you'll
- 1:04:13write your articles with AI, not
- 1:04:14necessarily—but your semantic
- 1:04:16research, keyword research, your
- 1:04:17brainstorming on topics, all that,
- 1:04:18you'll be doing with AI. So either you
- 1:04:22boost your margins, or you lower your
- 1:04:23prices compared to your competitors and
- 1:04:25you'll grow in volume. But it's really,
- 1:04:27really going to transform things today.
- 1:04:29And at the same time, big firms can't
- 1:04:31do that. BCG can't have all its slides
- 1:04:33done by an AI, because BCG—who
- 1:04:34tomorrow stops hiring, who loses a bit
- 1:04:36of quality, you lose a bit of visual
- 1:04:37quality...We, our slides, they aren't
- 1:04:39all the same, for example. But who said
- 1:04:41that consulting firm slides had to be
- 1:04:43...that deck 1 for client 1 must be
- 1:04:44exactly the same as deck 2 for client 2
- 1:04:46? It's huge. Our clients couldn't care
- 1:04:49less. They say it's beautiful, it's
- 1:04:51structured, it's practical; whether the
- 1:04:53logo is here or if for client 2 it's at
- 1:04:55the bottom, we don't care. So we waste
- 1:04:58zero time with that. I used to invest a
- 1:04:59bit of time, I spent my time putting
- 1:05:01the logo in the right place, fixing the
- 1:05:03slides, making drawings and such. Oh
- 1:05:06yes, the slides were beautiful in the
- 1:05:07end, but it cost the time of a junior
- 1:05:09being paid, they're paid € 10,000 a
- 1:05:10month there, or € 7,000 a month, and
- 1:05:12you charge that to your client. So
- 1:05:15a huge, huge disruption for consulting
- 1:05:17firms and service companies in general,
- 1:05:18where we will indeed be able to do much
- 1:05:20more, in my opinion, thanks to AI tools
- 1:05:21.
- 1:05:22I want to jump off that point about
- 1:05:24tools. You mentioned earlier that when
- 1:05:27you leave any kind of meeting, you take
- 1:05:29a voice note and say, "Here's what I
- 1:05:31saw." Quick, quick, quick, quick, quick
- 1:05:33, quick. "Write me an email, do this,
- 1:05:35do that." Uh, what makes you choose to
- 1:05:38use your voice note rather than, for
- 1:05:41example, putting a recorder on the
- 1:05:42table, recording the entire meeting,
- 1:05:45and then using an automated API that
- 1:05:47sends it to ChatGPT, then to Slack, and
- 1:05:49using automations to do it all? My
- 1:05:53underlying question is, how do you
- 1:05:55ultimately choose your daily workflow
- 1:05:57given the plurality of tools we're
- 1:05:59facing? That's a real issue. There are
- 1:06:02maybe 100 AI tools coming out a day
- 1:06:05from companies wanting to launch a SaaS
- 1:06:07, a thing, a gadget, each time with
- 1:06:09completely different or sometimes
- 1:06:11similar features. You can't even tell
- 1:06:14the difference between them, apart from
- 1:06:16the landing page. How do you choose?
- 1:06:19It's quite hard, you mustn't spread
- 1:06:20yourself too thin, that's for sure. Uh,
- 1:06:22you should always start with the
- 1:06:24generalists. Once again, dig into
- 1:06:26ChatGPT or Claude. Pick one or the
- 1:06:28other, it doesn't matter, Claude is
- 1:06:29very good, but explore everything it
- 1:06:31can do. For me, that's my first piece
- 1:06:33of advice. Uh, next, what I look at,
- 1:06:35well, that also depends on B2B versus
- 1:06:37B2C. B2C, have fun, you have a fairly
- 1:06:40wide range of tools. In B2B, we are
- 1:06:42constrained by tool security, that's
- 1:06:44clear.
- 1:06:44I would never put an unknown tool in
- 1:06:46the hands of a company. It's impossible
- 1:06:47. It's too dangerous. Tomorrow, if you
- 1:06:50want to make a competitor to, I don't
- 1:06:52know, what's a nice tool? Like Uta,
- 1:06:54which is my meeting tool. You want to
- 1:06:56create a Uta just as good, just as
- 1:06:57powerful? If you're a 13-year-old in
- 1:07:00India on your couch in Bangalore, you
- 1:07:02can create it thanks to AI. In fact,
- 1:07:04you'll create the same tool. So it's
- 1:07:06very dangerous for companies to be
- 1:07:07faced with many tools that are
- 1:07:08beautiful, powerful, and look serious.
- 1:07:10Some are French companies, others are
- 1:07:1213-year-old kids on their couch in
- 1:07:14their underwear in Bangalore. So there
- 1:07:16is one you want to trust more than the
- 1:07:18other. Except that companies don't know
- 1:07:19who is behind the tools, who is
- 1:07:20actually behind them. So we spend a lot
- 1:07:21of time telling them, "OK, let's take
- 1:07:23the example of meeting minutes, it's a
- 1:07:24need you have, you need to implement a
- 1:07:25tool. Well, we are going to screen all
- 1:07:27the tools for you and see which one is
- 1:07:29really secure because be careful not to
- 1:07:30send all your data to things that
- 1:07:31aren't secure. So, in business, we do
- 1:07:34it like this: we look at all the use
- 1:07:36cases, we tell ourselves there are 6 or
- 1:07:377 tools that might be interesting for
- 1:07:39you, we look at the budget, and
- 1:07:40depending on the budget, we say:" Well,
- 1:07:41we'll have to focus on two or three
- 1:07:43tools, or at least two or three
- 1:07:44categories of tools, "and for those, we
- 1:07:46pick the most secure one for you.
- 1:07:48That's a bit of our B2B approach. In
- 1:07:51B2C, obviously, you are a bit freer
- 1:07:52because you're dealing with less
- 1:07:54confidential information and so on.
- 1:07:56Well, on that note, I would rather tell
- 1:07:58you to keep yourself a bit updated on
- 1:07:59the latest tools that come out through
- 1:08:01newsletters; in Misteriad, we often
- 1:08:02talk about the latest tools. Now, there
- 1:08:04are some random lists that I advise
- 1:08:05against. They make lists of tools that
- 1:08:07—well, not us, but people make lists
- 1:08:09of tools that aren't very useful. Look
- 1:08:11instead at the tools that are raising
- 1:08:12money, it's that simple. But for me,
- 1:08:14it's a topic; when I see a great slide
- 1:08:17generation tool, I go and think, should
- 1:08:20I use it or not? I'll type into the
- 1:08:23internet what are the best AI slide
- 1:08:24generation tools. I look at those who
- 1:08:25have raised money. Well, that means
- 1:08:27they're more serious, that they're here
- 1:08:28to last, and so on. I try to avoid the
- 1:08:29little unknown Chinese tool. You see,
- 1:08:31with James Spark, we hesitated a lot at
- 1:08:32first. We saw that it was serious.
- 1:08:33There are big companies using it. We
- 1:08:35saw that they had raised a lot of money
- 1:08:37. Well, the probability that everyone
- 1:08:38is wrong is also a bit low. So for us,
- 1:08:40I look at it a lot like that, and
- 1:08:42indeed, pick two tools, three tools
- 1:08:43maximum. Look at where you have the
- 1:08:45most needs. You see, for example, for
- 1:08:47me, reports, I do them with ChatGPT,
- 1:08:49but reports because I do few reports
- 1:08:51where I can put a card like PL. So I
- 1:08:53don't use that, and I don't use Slack,
- 1:08:56I don't need it to send automatic
- 1:08:58emails and so on. So in this case, I
- 1:09:01tell myself my use case is more about
- 1:09:03making little voice notes and getting
- 1:09:04short reports that I send via
- 1:09:05copy-paste directly. So, I look at the
- 1:09:08use case like that, and depending on it
- 1:09:09, I pick two or three tools, but that's
- 1:09:11enough. Perplexity stays free, stay on
- 1:09:14free by the way, for everything you can
- 1:09:16use for free, keep Perplexity free. I
- 1:09:18would take a paid ChatGPT, a free
- 1:09:19Perplexity, maybe a paid James Spark,
- 1:09:21it's still a super powerful tool. James
- 1:09:23Spark or the equivalent, Manus, it's
- 1:09:25kind of the future of AI, so it lets
- 1:09:26you see something. When you already
- 1:09:29have those three, 99%of the whole world
- 1:09:31doesn't use any of the three, right? So
- 1:09:34already with those three, that's 3 x 20
- 1:09:35€ a month, or even Perplexity is free
- 1:09:37, so more like 2 x 20 € a month.
- 1:09:39You're already doing a lot of things,
- 1:09:41completely. You're talking about
- 1:09:42ChatGPT where you do your little
- 1:09:43reports. How do you use it yourself? Is
- 1:09:46it ChatGPT, just a new chat, or did you
- 1:09:48create a GPT or an AI agent? I've
- 1:09:50created few GPTs, or none at all on my
- 1:09:52end. Then again, I have a somewhat
- 1:09:55special job, where I'm in the field
- 1:09:58almost all day. I'm giving conferences
- 1:10:00all day, so I have fewer office use
- 1:10:01cases, so I haven't created any GPTs,
- 1:10:03well, I've created plenty, but I don't
- 1:10:05use them daily.
- 1:10:06For those listening, can you explain
- 1:10:08what a GPT is first? a GPT, sorry. The
- 1:10:10problem with ChatGPT is that every time
- 1:10:13I want to use it, I have to give it a
- 1:10:15new instruction, a new prompt; it
- 1:10:17doesn't remember me. Now, there is
- 1:10:19memory and so on, it's improving, but
- 1:10:21let's assume it starts from a blank
- 1:10:23slate. So, I might want to, for example
- 1:10:26—let's take a use case—I'm a real
- 1:10:28estate agent, I do viewings every day,
- 1:10:31and I have a visit report template that
- 1:10:34I want ChatGPT to follow. I don't want
- 1:10:36to explain the visit report template to
- 1:10:38it every single time. Plus, I only do
- 1:10:40viewings in English. So, I want my
- 1:10:42report to be translated into English. I
- 1:10:43don't want to explain how to do it and
- 1:10:45the template and the English every
- 1:10:46single time. So, I tell myself, I'll
- 1:10:48create a little robot that I teach to
- 1:10:49follow my template and speak in English
- 1:10:51. So, on ChatGPT, you can—it's called
- 1:10:53creating your own robot, creating your
- 1:10:55GPT—you give it an instruction. So,
- 1:10:57for me, the instruction would be to
- 1:10:58tell it that every time I trigger you,
- 1:11:00you will follow this interview template
- 1:11:01, you will respond in English, and so
- 1:11:03on. You can give it documents that it
- 1:11:06can use for inspiration. So I would
- 1:11:09tell it, look at the attachment, I've
- 1:11:10put in some English template documents
- 1:11:12that you can use as a reference. I give
- 1:11:13it that once and for all. I call my
- 1:11:15robot" Visit Report. "And now, every
- 1:11:18time I go to ChatGPT, instead of going
- 1:11:20back to ChatGPT and telling it," You
- 1:11:22are a visit expert, you will follow my
- 1:11:24template, etc., "I go to the robot I
- 1:11:26called" Visit Report "and there, I just
- 1:11:28tell it," I just visited a property,
- 1:11:30etc., "and automatically, it remembers
- 1:11:32that it has to follow my template, use
- 1:11:34my documents for reference, and do it
- 1:11:36in English. And so, this way, you can
- 1:11:38provide what are called
- 1:11:39pre-instructions, reference documents
- 1:11:40that you don't need to give it every
- 1:11:42time. That works well. We see many HR
- 1:11:44chatbots in companies. Actually, they
- 1:11:46put all their HR procedures behind a
- 1:11:48robot. They publish it on ChatGPT. The
- 1:11:51robot is called" HR Assistant "and as
- 1:11:52soon as someone joins the company, they
- 1:11:54talk to the HR Assistant. They say,"
- 1:11:55When can I take my vacation? "When do I
- 1:11:57do this?" And actually, the HR
- 1:11:58assistant on ChatGPT, well, it will
- 1:12:00query the famous document that the HR
- 1:12:02teams gave it 2 years ago, 2 months ago
- 1:12:03, and uh that allows for providing
- 1:12:05pre-instructions that way. So there you
- 1:12:07go, that's a bot, it works well, you
- 1:12:08can make many GPTs. But me, I don't
- 1:12:11really, I don't really have any...and
- 1:12:13yet, I think I query GPT 30 times a day
- 1:12:16, 20, 30 times a day about all sorts of
- 1:12:18things. I still prepare my meetings
- 1:12:20with it.
- 1:12:22GPT/Perplexity, I prepare my meetings
- 1:12:24with that. I go see someone, I say give
- 1:12:26me the three current events there. Give
- 1:12:27me those three news items. Here is what
- 1:12:29I sold to a small competitor. Tell me a
- 1:12:30bit how, based on these news items.
- 1:12:34That, that helps me a lot. Preparing,
- 1:12:36uh, what do we call that? Sales pitches
- 1:12:39. I use it a little bit but not GPTs
- 1:12:41yet. I need to, I need to find one that
- 1:12:42would really help me. And it changes so
- 1:12:44much, actually. For my part, my job is
- 1:12:46also to test as many tools as possible
- 1:12:48constantly because the nightmare of a
- 1:12:50consultant is showing up to a client
- 1:12:51who knows more or knows a tool you
- 1:12:53don't know. When it's tools not known
- 1:12:55at all, it's fine, it's not a big deal.
- 1:12:57I tell him, listen, we don't open 100 a
- 1:12:59day. However, for the 10, 15, 20, 30
- 1:13:01that are quite serious, that are
- 1:13:03evolving a bit, I am constantly testing
- 1:13:05them.
- 1:13:05H
- 1:13:06which means I spend less time on
- 1:13:08ChatGPT for example and more on Gemini,
- 1:13:10Claude every day, I think I use a bit
- 1:13:12of Gemini, a bit of Manus, a bit of
- 1:13:14Claude, NTA every day, uh, and a little
- 1:13:16French one every day. So you see, I'm
- 1:13:18more into testing a bit of everything
- 1:13:20rather than having a good,
- 1:13:21well-calibrated process on ChatGPT.
- 1:13:23You're right to point out this, the
- 1:13:24evolution of the tools, which is
- 1:13:26ultra-fast. Uh, already when we see the
- 1:13:28evolution of ChatGPT in 2 years, it's
- 1:13:30just, uh, phenomenal. Before, we
- 1:13:31couldn't even make, make, make images.
- 1:13:33Uh, ChatGPT, the, the very first one,
- 1:13:35what was it? Was it 2.5 or or 3.5
- 1:13:383.5.5.5
- 1:13:40it was pretty crazy. There have been
- 1:13:41several. There have been a dozen
- 1:13:42versions in all. Now only 5 is
- 1:13:44accessible. The evolution of usage is
- 1:13:47insane. I think we’re only at the
- 1:13:48very beginning compared to everything
- 1:13:50that's coming. And, uh, that multiplied
- 1:13:52by, uh, hundreds of tools. Uh, I say
- 1:13:55hundreds. I could say tens of thousands
- 1:13:57, but let’s say hundreds for those
- 1:13:59that are more or less serious, a few
- 1:14:01dozen that are really very serious, and
- 1:14:02then about fifteen, like you said, uh,
- 1:14:04that we can, we can trust and really
- 1:14:05build and move forward with. How do you
- 1:14:08manage to stay constantly up to date
- 1:14:10while still growing a hyper-growth
- 1:14:12company? Because it requires
- 1:14:13competitive intelligence on one hand
- 1:14:14regarding competitors, and on the other
- 1:14:16hand regarding tools, all while staying
- 1:14:18focused on the growth of a company
- 1:14:19that’s doubling every year. How do
- 1:14:21you do it?
- 1:14:22It’s hard, it’s very hard. I spend
- 1:14:23an hour a day staying informed.
- 1:14:24OK.
- 1:14:25One hour a day of staying informed.
- 1:14:26Actually, I force myself to publish,
- 1:14:28which is a pretty good trick. Since I
- 1:14:29force myself to post on LinkedIn every
- 1:14:31morning, I have no choice but to stay
- 1:14:32informed because if I don’t do my
- 1:14:33research, I have nothing to say, so to
- 1:14:35speak. So, uh, I do my research every
- 1:14:36morning from 8 to 9 AM, uh, on AI, and
- 1:14:38I’m on Twitter. I’m big on Twitter,
- 1:14:41Twitter plus a newsletter I like, an
- 1:14:43American one called The Rundown. It’s
- 1:14:46the English equivalent of Mysteria, or
- 1:14:47Mysteria is the French equivalent of
- 1:14:48The Rundown. Put it that way. So, I do
- 1:14:51my research like that for about an hour
- 1:14:54a day, not so much on the tools, by the
- 1:14:56way. I do it a little bit on tools, but
- 1:14:58much more on current events. You see,
- 1:15:00it’s important for me to know, is
- 1:15:01OpenAI developing other things? Did
- 1:15:04Mistral partner with Orange? If so,
- 1:15:06what does that mean? It wasn’t them
- 1:15:09who did it, by the way, it was more
- 1:15:11OpenAI, but you see, I’m trying more
- 1:15:12to understand the macro ecosystem, how
- 1:15:14it evolves within companies. So, I do a
- 1:15:16lot of research on that for my part.
- 1:15:19Plus, we have internal briefings every
- 1:15:20week; all Mysteria trainers and
- 1:15:21consultants are trained by an external
- 1:15:23person who comes to brief us. So, I
- 1:15:25participate in that too, and I listen
- 1:15:26to it. Uh, so that’s the research we
- 1:15:28do internally, which means there’s
- 1:15:29little chance there’s an AI topic I
- 1:15:31don’t master today, you see, it’s
- 1:15:32impossible. Uh, for someone whose
- 1:15:34full-time business isn't this—I don't
- 1:15:36have a choice, it is my full-time
- 1:15:38business and I love doing it. But at a
- 1:15:40minimum, read a newsletter, the one
- 1:15:42from Mysteria, it’s once a week,
- 1:15:44there are five news items, one news
- 1:15:46item—
- 1:15:48if you read that, you know everything
- 1:15:50about, well, you know everything.
- 1:15:51You're on top of a little bit of what's
- 1:15:52happening in AI and testing one small
- 1:15:54tool a week, you know, over the weekend
- 1:15:55, I test a small tool, that's something
- 1:15:57people can do quite easily. Uh, just
- 1:15:59with that, same thing, with that you'll
- 1:16:01be better than 95%of your competitors,
- 1:16:02and it's better to test one or two
- 1:16:04tools and actually deploy them all the
- 1:16:05way. Your competitors aren't doing it,
- 1:16:07so do it, rather than telling yourself
- 1:16:08you'll test two or three a week and
- 1:16:09then in the end, none of them go into
- 1:16:11production. So for me, I'd say read a
- 1:16:13newsletter, Mystery AI or another, and
- 1:16:15uh, and test a tool every, every week,
- 1:16:17a small tool, and you'll already see a
- 1:16:19lot of things. By the way, in The
- 1:16:22Rundown, it's very good because The
- 1:16:23Rundown is a great newsletter—it's in
- 1:16:25English, mind you—but it gives you
- 1:16:26the three or four news items, they
- 1:16:27provide a few business use cases, and
- 1:16:29they describe one tool in a bit more
- 1:16:30depth, saying: "Here is what it does,
- 1:16:31here is what companies can do with it."
- 1:16:34So it's not bad, honestly, just that
- 1:16:36alone is good, and you have to be
- 1:16:37disciplined for those who love AI but
- 1:16:38whose business it isn't, you have to be
- 1:16:40disciplined. not to dig into everything
- 1:16:42, because we see everything, it's
- 1:16:44infinite, and we see everything and
- 1:16:46anything, and
- 1:16:47it goes in every direction, so you
- 1:16:48shouldn't, no, at first we dug into
- 1:16:50everything, we dug into everything, but
- 1:16:51after a while, you just want to jump
- 1:16:52out the window. So you have to be
- 1:16:53disciplined a bit, tell yourself there
- 1:16:55are lots of things in AI, I follow
- 1:16:56Mysteria or Rundown which are very
- 1:16:57serious, I see a bit of what's being
- 1:16:58done, and that'll already be pretty
- 1:17:00good.
- 1:17:00You have mentioned confidentiality and
- 1:17:02data several times, we could even talk
- 1:17:04about GDPR. Uh,
- 1:17:06what about that? Because sometimes some
- 1:17:08people, as we see, say, "No, but AI, we
- 1:17:10don't know what they'll do with our
- 1:17:12data, we could lose everything, it's
- 1:17:14too dangerous," et cetera. "Others say,
- 1:17:17" No, not at all, when you just get the
- 1:17:19subscription, well, you click this box
- 1:17:21and then it's completely secure. "You
- 1:17:23who implement this daily in companies,
- 1:17:25with entrepreneurs, with individuals,
- 1:17:27what is the truth? There is a big issue
- 1:17:30, which is that most AI tools are
- 1:17:32American and they will remain American.
- 1:17:34I don't know if you've seen the latest
- 1:17:36funding rounds, but OpenAI is set to
- 1:17:37raise nearly 100 billion at a 500
- 1:17:39billion valuation. We are talking about
- 1:17:41a private company that will be worth
- 1:17:42500 billion, which had no product 2 or
- 1:17:443 years ago. So that is the first point
- 1:17:46. Their competitor, Anthropic, just
- 1:17:49raised 18 billion or 14 billion at a
- 1:17:51200 billion valuation. So we will not
- 1:17:53catch up with those people. You have to
- 1:17:54forget about Mistral, meanwhile, the
- 1:17:56French one just raised 1 billion at a
- 1:17:5714 billion valuation. 14 versus 180
- 1:17:59versus 500. Americans, as soon as they
- 1:18:01raise their hand, they find 50 or 100
- 1:18:03billion wherever they want, all over
- 1:18:04the world. So I think we won't catch up
- 1:18:06with that. So the big players in AI are
- 1:18:08American and will remain so. So there
- 1:18:11is an issue, which is that the data you
- 1:18:13put into those tools ends up in the
- 1:18:15United States. That much is clear. At
- 1:18:17the very least, because OpenAI is
- 1:18:18opening servers, notably in Europe. At
- 1:18:20the very least, it might perhaps stay
- 1:18:21in Europe with Microsoft Copilot. It
- 1:18:22could also stay in Europe. But on
- 1:18:25servers that belong to Americans. And
- 1:18:27it's quite funny because, very often
- 1:18:29with AI, we hear a lot that it will
- 1:18:31feed the models, it will train...if all
- 1:18:33the data we put in goes to feed the
- 1:18:35American models, etc. And we have many
- 1:18:37clients who tell us," Oh no, I don't
- 1:18:39want to use ChatGPT, it's the big bad
- 1:18:42wolf. "However, I use SharePoint, folks
- 1:18:44. SharePoint, those are servers that
- 1:18:46belong to Microsoft, the same way that
- 1:18:48OpenAI servers are Microsoft servers.
- 1:18:51The problem, however, behind all that,
- 1:18:52is that the Microsoft server is subject
- 1:18:54to the American Cloud Act. The American
- 1:18:56Cloud Act, which explains to us that,
- 1:18:58in fact, the Americans, it is the
- 1:18:59extraterritoriality of American law. At
- 1:19:01any time, the U.S. government or the
- 1:19:04CIA, or at least major U.S. public
- 1:19:06agencies, can get their hands on 100%of
- 1:19:08the data belonging to companies that is
- 1:19:11stored by American firms anywhere in
- 1:19:13the world. It means a Microsoft server,
- 1:19:15whether it's in Mississippi, Saint-
- 1:19:18Étienne, Toulouse, or Poland—if it's
- 1:19:20a Microsoft server, the Americans can
- 1:19:22get their hands on the data at any time
- 1:19:25. So we often say, okay, for businesses
- 1:19:27, we'll use secure tools that don't
- 1:19:29train their models with your data. You
- 1:19:31have to be careful with that; it's a
- 1:19:32shame. Be careful with the free
- 1:19:34versions of these tools, by the way.
- 1:19:35Everything you put on the free cloud,
- 1:19:37free GPT, free Perplexity—they take
- 1:19:38it and train their models with it. The
- 1:19:41probability that you'll find what was
- 1:19:43said tomorrow in a trained model is
- 1:19:44virtually zero. But still, I advise
- 1:19:46against doing that, and we never put
- 1:19:48companies on these free versions. Once
- 1:19:50you've moved to the paid versions of
- 1:19:52these tools, they comply with GDPR and
- 1:19:55they do not train on your data. When
- 1:19:58you're on a paid version—I'll take
- 1:19:59ChatGPT, which is supposedly the least
- 1:20:01secure. When you're on the paid version
- 1:20:03, they don't train their models with
- 1:20:05your data. That's the first thing. It's
- 1:20:08GDPR compliant, so you can delete your
- 1:20:10data and so on. However, yes, it will
- 1:20:12pass through American servers, but
- 1:20:14you're already routing everything
- 1:20:16through American servers anyway. So I
- 1:20:18often say: it's a fake problem that
- 1:20:19we're trying to invent for ourselves
- 1:20:21today. But we probably should have
- 1:20:22asked ourselves that question 10 years
- 1:20:23ago when the cloud was born, because
- 1:20:24the real problem is that we don't have
- 1:20:25a French cloud. And that's the real
- 1:20:28issue: if we had a French cloud, then
- 1:20:29we could say," Okay, OpenAI, as long as
- 1:20:31their cloud isn't with a French
- 1:20:32provider, as long as they aren't using
- 1:20:34OVH—I'll put it that way—then okay,
- 1:20:35we won't use them. "" But we don't have
- 1:20:38a French cloud; we have OVH, but it's
- 1:20:40tiny compared to giants like Microsoft,
- 1:20:42Amazon, and so on. "And so today we are
- 1:20:44forced to use things where everything
- 1:20:46we use daily is stored on servers that
- 1:20:48belong to Americans, and unfortunately,
- 1:20:49that won't change. So, be careful, yes.
- 1:20:52Don't just use any tool; yes, choose
- 1:20:54paid corporate tools rather than free
- 1:20:55ones to prevent them from training
- 1:20:57their models. Yes. But having a general
- 1:21:00psychosis that says we can only use
- 1:21:02Mistral because of the data problem. No
- 1:21:05, if you want to use Mistral for
- 1:21:06sovereignty, because you want to
- 1:21:08promote French companies, okay, that's
- 1:21:10fine, and we should all be a bit more
- 1:21:12patriotic like that, why not. But the
- 1:21:14data argument, you have to be a bit
- 1:21:16careful because data—I've seen
- 1:21:17executive committees tell me they would
- 1:21:18never deploy ChatGPT. And during the
- 1:21:20meeting, she gives her colleague
- 1:21:21something and says," Wait, here, I sent
- 1:21:22you the report via WhatsApp. "" But
- 1:21:24what is WhatsApp? "If it's not an
- 1:21:26insecure tool, 100%of it goes to
- 1:21:28American servers, and it's actually
- 1:21:30stored on those servers. That's not
- 1:21:32ChatGPT at all." So you see, there's a
- 1:21:34big, uh, there's a lot of ignorance
- 1:21:35about this. "It's normal; we haven't
- 1:21:37been very well sensitized to it. But
- 1:21:38when you're not sensitized, rather than
- 1:21:40having grand principles, just say,"
- 1:21:41We'll be careful, we'll keep an eye on
- 1:21:43things, and we won't do anything too
- 1:21:44stupid. "" But the grand principles of'
- 1:21:46We absolutely forbid ChatGPT and the
- 1:21:48like '. "On the other hand, we talk to
- 1:21:50each other on WhatsApp at work; that's
- 1:21:52not very consistent, completely. Uh,
- 1:21:54we've talked a lot about usage in
- 1:21:56companies, large groups or others, uh,
- 1:21:58entrepreneurs, and then you mentioned
- 1:22:00individuals that you also train. Uh,
- 1:22:03what are the uses we can have in our
- 1:22:05personal lives? Whether it's for
- 1:22:07training, learning, maybe managing
- 1:22:09finances better, doing sports, why not.
- 1:22:12Uh, that could be interesting. You've
- 1:22:14already mentioned quite a few that work
- 1:22:15well. Uh, you can use it as a personal
- 1:22:17coach; I like that, I use it as a
- 1:22:19personal coach with a lot of
- 1:22:20perspective. Be careful, there have
- 1:22:22been issues with people who committed
- 1:22:24suicide because of unsecured AIs. So,
- 1:22:25first of all, when you do that, use
- 1:22:28mainstream AIs like ChatGPT. Meaning
- 1:22:30the day it pushes someone toward
- 1:22:32suicide, in a millisecond, uh, there
- 1:22:33are people monitoring everything said
- 1:22:35on ChatGPT, so they are highly watched,
- 1:22:37so they can't do much, whereas there
- 1:22:39are open-source AIs that can make you
- 1:22:40do anything. So, do this instead; I do
- 1:22:42it on ChatGPT, but I have a personal
- 1:22:44coach where I tell it everything,
- 1:22:45essentially, and again, everyone will
- 1:22:47do as they please, I don't know if it's
- 1:22:48the best practice. I give it everything
- 1:22:50. I operate on the principle that my
- 1:22:52health data is managed by Apple today.
- 1:22:53I give it all; I'm using an American
- 1:22:55service. It sends everything to my
- 1:22:56iPhone, which is with Apple. So, my
- 1:22:58health data is already with the
- 1:22:59Americans, as far as I'm concerned. I
- 1:23:00assume they have more to offer me in
- 1:23:01terms of positives than negatives by
- 1:23:03knowing all my data. So, in any case, I
- 1:23:06have a ChatGPT that knows everything
- 1:23:08about me and it's a personal coach.
- 1:23:10Meaning I talk to it two or three times
- 1:23:12a week about what I'm doing, what's
- 1:23:13going well, what's not, and so on. And
- 1:23:16instead of telling it," You're a
- 1:23:17psychologist, though I don't know if I
- 1:23:19need one, "I instead tell it to give me
- 1:23:20actions I can take to improve what you
- 1:23:22feel is lacking in me. Like, I told you
- 1:23:24for three weeks that I didn't have time
- 1:23:26to see my friends because I was working
- 1:23:27too much. Well, when I explain all that
- 1:23:28, without even being very specific, I
- 1:23:30just tell it," Here, I did this today,
- 1:23:31that, and so on and so forth. "Well, it
- 1:23:32will tell me,' Wait, I realize you
- 1:23:34haven't seen your friends in three
- 1:23:35weeks. So, send a message to two or
- 1:23:37three of your friends to see how they
- 1:23:38are; it will get you back in the loop,
- 1:23:40and you can book a breakfast. '" See,
- 1:23:41it's just really simple stuff. It helps
- 1:23:43me a lot with that. Uh, you see, a
- 1:23:44really simple thing about my personal
- 1:23:46life. Uh, I realize I work so hard
- 1:23:47during the week that I end up at 10
- 1:23:49p.m. on Friday, closing my laptop and
- 1:23:51saying, "What am I doing this weekend?"
- 1:23:52Except your friends don't wait until 10
- 1:23:54p.m. to decide what they're doing for
- 1:23:55the weekend. So you end up seeing the
- 1:23:56same people, doing the same things.
- 1:23:58Well, you talk about that with your AI
- 1:24:00for three weeks. Well, I wasn't even
- 1:24:01talking about that, I was talking about
- 1:24:03my life. And he finally told me, "Look,
- 1:24:05here's my advice: let's do one thing—
- 1:24:07take a weekend, sit down, and plan your
- 1:24:08weekends for the next six weeks, at
- 1:24:10least then you won't be in a rush
- 1:24:11looking for this or that.""It's so
- 1:24:12simple, but it's going to change my
- 1:24:13life." I mean, really, I've been doing
- 1:24:14so much better since I started doing
- 1:24:15that. So you see, I have a personal
- 1:24:17coach like that
- 1:24:18who helps me a bit with these types of
- 1:24:19things. Sometimes he asks me, "What
- 1:24:21would you like to improve about
- 1:24:23yourself as an entrepreneur?""Um, so I
- 1:24:25don't know, I tell him,' Well, here's
- 1:24:26what I do. '""He says,' Well, wait, I
- 1:24:27think you'd benefit from being more
- 1:24:29creative. '""So I say,' Well, give me
- 1:24:30some creativity exercises. '""And just
- 1:24:32recently, he gave me a creativity
- 1:24:33exercise that is brilliant." Every
- 1:24:35morning, he sends me—well, I trigger
- 1:24:37it, but he sends me a theme every
- 1:24:39morning, or basically, he gives me a
- 1:24:41theme where I have to come up with 10
- 1:24:43ideas. For example, he'll say, "Imagine
- 1:24:4610 features you could buy on Facebook."
- 1:24:49"The next day, he says,' Imagine 10
- 1:24:51companies that Snapchat could acquire.
- 1:24:52'""And why?""Imagine 10 Tesla models
- 1:24:56that Tesla could create and why?" It's
- 1:24:59a creativity exercise that is super
- 1:25:00interesting. He's the one who gave it
- 1:25:02to me because he told me, "In your job,
- 1:25:04you see the same people a lot, you give
- 1:25:06the same talks often, so your
- 1:25:07creativity is fading a bit.""That's a
- 1:25:09big topic for me, I love the personal
- 1:25:10coach." And obviously, you can input
- 1:25:12all your data, you know. Me, my health
- 1:25:13data, but everyone will do what they
- 1:25:15want, of course. I think that if there
- 1:25:17are some listening to us in the
- 1:25:18comments and so on, not everyone will
- 1:25:20agree, but it finds things in my blood
- 1:25:21tests that my doctor doesn't find. It's
- 1:25:24true. It's, and again, I'm not advising
- 1:25:27you to get treated based on ChatGPT,
- 1:25:28but telling it, "Here is my blood test,
- 1:25:30what do you think?" It knows my
- 1:25:32lifestyle, so very quickly it tells me,
- 1:25:34"Well, you're not getting enough sleep.
- 1:25:35""That's why this thing is high." My
- 1:25:37doctor doesn't. "So those are things
- 1:25:39that can be done." Now, there might be
- 1:25:41more secure AIs than this one to do it.
- 1:25:44Personally, I give it a lot of personas
- 1:25:45, and then obviously everything related
- 1:25:47to workout programs, I do that a lot.
- 1:25:50Uh, what is going to help me learn? Hey
- 1:25:52, I've got this topic. Oh, wait, I
- 1:25:54actually have one I've been using
- 1:25:55lately that I love. When I read the
- 1:25:56newspaper, every morning I read Les
- 1:25:58Échos, and there are plenty of things
- 1:25:59I don't understand. Screenshot, explain
- 1:26:01it to me in simple words. It's
- 1:26:02wonderful. It's really about things
- 1:26:04like, why does the increase in debt in
- 1:26:05such-and-such country have an impact on
- 1:26:07youth employment rates in this other
- 1:26:08country? They explained it quickly
- 1:26:10there, and I didn't get it. Snap a
- 1:26:11photo for GPT, and it explains it to me
- 1:26:13. Actually, you learn so much faster
- 1:26:15with that, and that's a purely personal
- 1:26:16thing, or one that works in
- 1:26:17professional life too, but it's a
- 1:26:18personal thing that works well. And
- 1:26:20another point, a final one that is less
- 1:26:22about learning, but regarding movies: I
- 1:26:25realized recently, I was watching a
- 1:26:27film, *Tenet*or I don't know what, that
- 1:26:29was a bit complex, and there was a
- 1:26:31scene I didn't understand. I tell it,
- 1:26:33the guy is referring to something, but
- 1:26:34I didn't understand what, and I ask
- 1:26:36ChatGPT, it explains it to me, and
- 1:26:37what's more, because it knows the movie
- 1:26:38, it has read a thousand forums where
- 1:26:40people explain it, and it unlocks it
- 1:26:41for you, and that's quite, quite cool.
- 1:26:43You watch the end of the movie and you
- 1:26:44understand why he did that, and there
- 1:26:45are tons of cases like that. There are
- 1:26:47indeed plenty of examples, and you
- 1:26:48mentioned the fact of analyzing blood
- 1:26:50tests, uh,
- 1:26:52in the context of health coaching, etc.
- 1:26:54, or the data that a Whoop or an Apple
- 1:26:55device, whatever, might give you. Those
- 1:26:57are good examples that I've used myself
- 1:26:59and that I also use on a daily basis. I
- 1:27:01use AI. However, there was a rather
- 1:27:04disappointing effect, at least
- 1:27:05perceived by people overall, during the
- 1:27:07release of the new version of ChatGPT,
- 1:27:09GPT-5, where they had mentioned the
- 1:27:11fact that, at a minimum, all answers to
- 1:27:13certain questions would be at a PhD
- 1:27:15level, it was going to be much more
- 1:27:17advanced, and so on. Personally, I
- 1:27:21wasn't necessarily disappointed by this
- 1:27:22new version. In your case, what do you
- 1:27:24think about it? Why was there that
- 1:27:26effect? Is it because it had evolved so
- 1:27:28much in previous times that we were
- 1:27:29expecting something incredible? That
- 1:27:31could be interesting.
- 1:27:32I think it's a good mix, indeed, of all
- 1:27:33that. In my opinion, the model is
- 1:27:35better. The model is better. There's no
- 1:27:37debate. That's for sure, benchmark-wise
- 1:27:39, the model is better.
- 1:27:40The model has evolved. Does it have a
- 1:27:42way of answering you that's a bit
- 1:27:44different? And actually, it's super
- 1:27:47hard for a human eye to say it's better
- 1:27:49or worse. What does a better model even
- 1:27:50mean? In fact, I have a lot of people
- 1:27:52telling me that since it answers
- 1:27:53differently, it's worse. Well, not
- 1:27:55necessarily. So it's not worse. It
- 1:27:58answers in a different way, that's for
- 1:27:59sure. There's a big bias too, as you
- 1:28:02said, since it evolved so much between
- 1:28:04versions 3 and 4, and between 4 and 1,
- 1:28:06that now the improvements are more
- 1:28:08marginal, that's certain. So we were
- 1:28:11expecting a huge jump. No, the jump is
- 1:28:13smaller, that's certain too. And then,
- 1:28:15we just get used to it. We get used to
- 1:28:17the model. Recently, I think it was GPT
- 1:28:19-4, there were these charts: they ask
- 1:28:21people all the time what they think
- 1:28:23about performance, and the performance
- 1:28:24curve of GPT-4 dropped in people's
- 1:28:26opinion between March '24 and the end
- 1:28:27of '24. Everyone thought, "That's weird
- 1:28:30, that's not normal." OpenAI said, "
- 1:28:31It's the same model, folks, it hasn't
- 1:28:33changed," so at the very least it's the
- 1:28:35same, it can't be worse. And in fact,
- 1:28:37when you looked deeper, scientists
- 1:28:39realized that as we got used to ChatGPT
- 1:28:41, we were less and less impressed by
- 1:28:43its answers, and so we felt it was
- 1:28:45worse because, back then, you'd tell it
- 1:28:47, "Give me a recipe." It would give you
- 1:28:49a recipe and you'd say, "Wow, it's
- 1:28:50incredible, it works so well." Six
- 1:28:53months later, on the same model, you
- 1:28:55wouldn't just ask for a recipe, you'd
- 1:28:56say, "X grams of this, X milligrams of
- 1:28:57that.""Look at the photo, look at my
- 1:29:00utensils, give me a recipe, milligram
- 1:29:02by milligram." And then, it wasn't
- 1:29:03perfect. So, you'd say, "Ah, it's worse
- 1:29:04than before." No, it's not worse. It's
- 1:29:06just that we've learned how to use it
- 1:29:07and we are increasingly demanding. So,
- 1:29:09I think with GPT-5, the expectations
- 1:29:11have also increased a lot. They hyped
- 1:29:12it up a bit by telling us, "It's going
- 1:29:14to be incredible, it's a model like no
- 1:29:15other." It turns out it is better, but
- 1:29:17not by a massive amount. And so, I
- 1:29:19think there was a bit of disappointment
- 1:29:20as well. But that's mainly it, it's
- 1:29:21very biased, you know. In fact, it's a
- 1:29:23model that is objectively better, that
- 1:29:24reasons better, that writes better,
- 1:29:26that has fewer writing biases in what
- 1:29:27it does. Uh so no, it's a better model
- 1:29:30and there too, people shouldn't
- 1:29:32overthink the model. It's not the model
- 1:29:34that does the work, it's your way of
- 1:29:36using it and the quality of the product
- 1:29:37. Those are two big productivity
- 1:29:39factors for an AI tool. The model today
- 1:29:42, I think 90%of people, if you don't
- 1:29:44write GPT-5, you run 4 EOS behind it or
- 1:29:46Llama 3 or Claude 3.5, they don't know
- 1:29:49who they're talking to. That's normal,
- 1:29:51these models are quite similar. So
- 1:29:53there you go, completely. I have three
- 1:29:55final questions for you, two more
- 1:29:57business-related. The first one is, you
- 1:29:59are two with your brother as partners.
- 1:30:02How do you divide the roles? Well,
- 1:30:05until today, he was very B2C, and I was
- 1:30:07very B2B.
- 1:30:08OK.
- 1:30:09So that was a division that is changing
- 1:30:10a bit in September because the B2B side
- 1:30:12is really exploding for us. So, we're
- 1:30:13both going to be on B2B. We'll split
- 1:30:15the accounts, actually. We'll split the
- 1:30:17accounts. There are two or three
- 1:30:19industries I'm very strong in, so I
- 1:30:20keep those accounts. He is more into
- 1:30:22real estate, actually. He handles real
- 1:30:23estate accounts very well, so he keeps
- 1:30:25the real estate accounts, so we're
- 1:30:26splitting the accounts among ourselves.
- 1:30:29What we call the Misteria partners,
- 1:30:30meaning our four true consultant
- 1:30:32partners plus the three associates—
- 1:30:33there's a third one who joined—the 7
- 1:30:35of us, we split the accounts, actually.
- 1:30:37So, it's more like that. Today, we also
- 1:30:39split the tasks a bit. We bought back
- 1:30:41…so our two former companies have
- 1:30:43been integrated into the business, we
- 1:30:45bought them back. So, he will manage
- 1:30:46those two media outlets in addition; I
- 1:30:48don't manage them at all, I manage the
- 1:30:50recruitment parts a bit more, he
- 1:30:51doesn't do finance at all. We are
- 1:30:53really both splitting things up a bit
- 1:30:54like that, but we try not to do the
- 1:30:56same things. There are few subjects
- 1:30:57that we work on together.
- 1:30:58Which means that, in fact, I trust him
- 1:31:00100%, he trusts me 100%, and that makes
- 1:31:02things go super fast. So that's good.
- 1:31:05Completely. And that answers the other
- 1:31:08question a bit regarding the breakdown
- 1:31:10between B2C and B2B, which is a real
- 1:31:13topic. Many companies say you have to
- 1:31:15stay focused and that you can't fight
- 1:31:17three battles at once, or even two. Yet
- 1:31:20today, that is what you are doing, with
- 1:31:22major accounts on one side, more
- 1:31:23traditional companies on the other, and
- 1:31:25B2C as well. What made you decide to
- 1:31:28take on these three different fronts?
- 1:31:30Is it because there is a global market
- 1:31:32opportunity, while at the same time
- 1:31:34responding to the high demand from
- 1:31:36companies, as you mentioned, by
- 1:31:38recentering resources around B2B? That
- 1:31:41could be interesting.
- 1:31:41I would have given the same advice at
- 1:31:43the beginning, saying don't diversify,
- 1:31:45it's not a good idea. It turns out the
- 1:31:48AI market is so enormous that you take
- 1:31:50a little less risk when you diversify
- 1:31:52in that sector. That's for sure. There
- 1:31:55is also the fact that the two talk to
- 1:31:56each other a lot. Actually, we have a
- 1:31:58huge number of trainees and people
- 1:31:59trained in B2C who come back and say, "
- 1:32:01Okay, now we're going to train my
- 1:32:02company." So for us, that is very
- 1:32:04important. And conversely, I have many
- 1:32:06companies tell me, "We're an SME, we
- 1:32:08can't afford Misteria training."
- 1:32:10However, everyone has a CPF account, so
- 1:32:12come to us and we'll fund it, and
- 1:32:13anyone who wants to be trained can do
- 1:32:15so, because the employer doesn't have
- 1:32:16the right to force you to use your CPF.
- 1:32:18CPF is personal. However, the employer
- 1:32:20can very well say, "Listen, I don't
- 1:32:22have the budget to train you, sorry.""
- 1:32:23If anyone really wants to get trained,
- 1:32:25well, I'll ask Misteria to run a
- 1:32:26session for you.""You get a group of 5
- 1:32:28together, they come here to the company
- 1:32:29, we customize it, and so on," and
- 1:32:30there, they use their CPF. So B2C and
- 1:32:33B2B talk to each other a lot at our
- 1:32:35company, which is why we keep both. We
- 1:32:38brought someone on very quickly;
- 1:32:39someone arrived recently. In our
- 1:32:40company, they manage B2C, which is also
- 1:32:41important for us. I didn't want the two
- 1:32:43founders to be spread too thin between
- 1:32:45the two. We really have someone
- 1:32:46dedicated to that. And there you go,
- 1:32:49and the final point I would make, it
- 1:32:51was also a way for us not to put all
- 1:32:53our eggs in one basket in an ecosystem
- 1:32:56that is certainly huge, but where we
- 1:32:58still have little visibility on AI.
- 1:33:00It's it's always the same, even if I am
- 1:33:03100%convinced of the depth of the tech,
- 1:33:05we don't know how the market will
- 1:33:07evolve in 5 years. When you build a
- 1:33:09SaaS on a classic theme, well, you have
- 1:33:11a bit of visibility and so on, and
- 1:33:13there indeed you wonder, should I do
- 1:33:15B2B or B2C? For us, it's very, very
- 1:33:17hard. So we said let's put some eggs in
- 1:33:18several baskets, let's grow like that,
- 1:33:20let's try to build as many bridges as
- 1:33:22possible between B2B and B2C so that,
- 1:33:23as you said, it's not too scattered.
- 1:33:25And there you go, that was the choice,
- 1:33:26and it happened little by little. It
- 1:33:28wasn't a choice from the beginning, by
- 1:33:29the way, it was opportunities here and
- 1:33:30there. But right now, for instance, we
- 1:33:32are thinking about partnering with
- 1:33:33companies that only do B2C, saying
- 1:33:35maybe Misteria will do a big B2B pole
- 1:33:36and behind that we will have a partner
- 1:33:38who will handle the whole B2C side. We
- 1:33:40are still thinking about that. But yeah
- 1:33:41, that's the big argument behind it.
- 1:33:44Thank you Martin for everything that
- 1:33:46was discussed in this episode. It's a
- 1:33:48vast subject, a big subject, and also
- 1:33:50super interesting to see how things
- 1:33:52grow quickly. Looking forward to what's
- 1:33:54next. Congratulations on what you have
- 1:33:56built in just 2 years and for
- 1:33:58responding to this wave of opportunity.
- 1:34:00Uh, I loved this episode. And if you
- 1:34:02enjoyed this episode as much as I
- 1:34:04enjoyed hosting it, let us know as
- 1:34:06always by sharing this video, leaving a
- 1:34:08comment, or liking it. And then, if you
- 1:34:11are on a streaming platform, your
- 1:34:13favorite streaming platform, don't
- 1:34:14forget to leave 5 stars on Apple
- 1:34:16Podcast or here. Martin, I have one
- 1:34:18last question for you that I ask all
- 1:34:19our guests. Can you share the turning
- 1:34:21point that made all the difference for
- 1:34:23you, whether personally or
- 1:34:25professionally, that you haven't shared
- 1:34:27in this episode or perhaps never in any
- 1:34:29interview? It could be a simple
- 1:34:30sentence, an anecdote, something
- 1:34:32positive or less positive. You have
- 1:34:33free rein for your final words. Not
- 1:34:36easy. Uh, a turning point that changed
- 1:34:39a lot of things. Hmm. I'll have to
- 1:34:44think a bit in terms of
- 1:34:45entrepreneurship. What could have made
- 1:34:47me take the leap? Yes, maybe I did have
- 1:34:51one. I went to the United States, to
- 1:34:53Texas. I went to Texas. So, I I come
- 1:34:57from I come from Saint-Étienne. I'm
- 1:34:59not really I'm not Parisian at all, and
- 1:35:02I think that, especially outside of
- 1:35:04Paris, French people don't stop.
- 1:35:07Everyone says we're not good enough.
- 1:35:08Well, no, that's not it. Actually,
- 1:35:11teachers at school—I’m caricaturing
- 1:35:12it a bit—I find don't give much hope.
- 1:35:14You see, they tell you you'll have a
- 1:35:16small job and that it's just fine. They
- 1:35:19constantly tell you that bosses are
- 1:35:20stupid, they keep saying it. I think
- 1:35:21everything is just a bit negative. And
- 1:35:23when you come from that environment,
- 1:35:24honestly, you don't want to be an
- 1:35:25entrepreneur, you don't want to get
- 1:35:26started. They keep talking about people
- 1:35:29who started companies and crashed, and
- 1:35:30it's seen as shameful. They always talk
- 1:35:32about the guy in your high school
- 1:35:34making videos, and it's seen as
- 1:35:35shameful. That's the feeling, we're
- 1:35:37kind of stuck in that, you know. Paris
- 1:35:38is a little less like that, but
- 1:35:39otherwise in France, it's a bit gloomy.
- 1:35:42And the turning point was when I went
- 1:35:44to the US, I lived in Austin for a year
- 1:35:46, and you actually see the polar
- 1:35:47opposite where you can do whatever you
- 1:35:49want. I get to school, and everyone has
- 1:35:52started a company. When I say a company
- 1:35:54, it’s small Instagram pages with
- 1:35:5510,000 followers sometimes, or 5,000,
- 1:35:57food pages, you know. But everyone has
- 1:35:59a project in class; the teacher speaks,
- 1:36:01and 100%of the students raise their
- 1:36:02hands to participate, whereas in France
- 1:36:04, it’s embarrassing to be the one who
- 1:36:06participates. And you actually feel a
- 1:36:09population that doesn't ask questions,
- 1:36:10that wants to move forward, wants to do
- 1:36:12things, and isn't there thinking: "Yes,
- 1:36:14but is it really fair if I make money
- 1:36:16while my neighbor earns a little less?"
- 1:36:18No, everyone is pulled upward, you see.
- 1:36:20And that really shocked me a bit. And
- 1:36:22when I returned to France, I told
- 1:36:23myself, that's also why I wanted to
- 1:36:24start a business, I think. I told
- 1:36:25myself, that's what I want to do. I
- 1:36:26want to be around ambition. I want to
- 1:36:28be with people who want to change
- 1:36:30things a bit, who aren't ashamed to
- 1:36:31talk about money, who don't ask
- 1:36:33themselves a thousand questions, you
- 1:36:35know, who talk to clients, who take
- 1:36:37part, and so on. So that was a bit of a
- 1:36:38turning point. If I had to give one
- 1:36:40piece of advice, I'd tell everyone to
- 1:36:41go live in the US for a year. There are
- 1:36:42plenty of flaws in that country, but
- 1:36:44they have one quality: they ask
- 1:36:45themselves fewer questions before doing
- 1:36:46things. And I think if we asked
- 1:36:47ourselves fewer questions, if we all
- 1:36:48took ourselves a little less seriously,
- 1:36:50we'd get more done. And that's been a
- 1:36:52big topic, and it bridges the gap with
- 1:36:54school. So I was at HEC, and one day I
- 1:36:56heard someone at the Lyon campus say
- 1:36:58this. Everyone, someone asks in a
- 1:37:00podcast, "What's the big advantage of
- 1:37:02HEC?" Honestly, you never go to class,
- 1:37:04you just party. So well, there's no
- 1:37:07academic advantage, you don't really
- 1:37:09learn much. Sorry if there are any HEC
- 1:37:11people listening, but you don't learn
- 1:37:13anything. You meet cool people, that's
- 1:37:15for sure. Everyone tells you, "The
- 1:37:17network, the network." Yes, but I don't
- 1:37:19know if it's a game changer. However,
- 1:37:21he was saying there is one strong point
- 1:37:23: you're surrounded by ambitious people
- 1:37:24every day, and you're surrounded by
- 1:37:26teachers and a school that makes you—
- 1:37:28I was going to say makes you believe,
- 1:37:29no, makes you realize that anything is
- 1:37:31possible. And without it being BS, they
- 1:37:33tell you, "Oh, you want to start a
- 1:37:34company that will...? Well yes, we have
- 1:37:35students who've done it, so yeah, you
- 1:37:36can do it.""Oh, you want to start a
- 1:37:37company and have employees in the US?""
- 1:37:39Yeah, go ahead, we'll help you with
- 1:37:40some things." And that was a big
- 1:37:41turning point too, that you're in a
- 1:37:43school, an ecosystem where people don't
- 1:37:45put barriers on themselves, you know.
- 1:37:47Whereas I have friends who, you know,
- 1:37:49see things and say, "Damn, that guy has
- 1:37:51a Ferrari, damn, I'll never have that."
- 1:37:54There isn't a single thing in life
- 1:37:55where I tell myself, "I'll never have
- 1:37:56that or I'll never do that." Being
- 1:37:57president, maybe I'll do it, you know,
- 1:37:58I have no idea. In any case, it's
- 1:37:59possible. You say that everything is
- 1:38:01possible and that's what Americans do;
- 1:38:03they say everything is possible for
- 1:38:04them. What a school like a top-tier one
- 1:38:06brings a lot is ambition; they tell you
- 1:38:08that everything is possible, which is
- 1:38:09something we don't instill in young
- 1:38:11people. We tell them, "Oh no, don't
- 1:38:12make too much noise, don't speak too
- 1:38:13loudly." And you see, he has a big
- 1:38:14house. Yeah, but that doesn't interest
- 1:38:16us; we'll never be like that. You see,
- 1:38:17he studied in Paris. Yeah, but you
- 1:38:19don't have that luck, you. People are
- 1:38:21very much like that in France. They
- 1:38:22talk to their kids, their students,
- 1:38:23like that a lot. And, well, I think
- 1:38:25it's a real breakthrough to understand
- 1:38:27that, in fact, everything is possible.
- 1:38:29Yes, building a consulting firm worth
- 1:38:3010 billion is possible. It will be very
- 1:38:31, very, very hard. The probability of
- 1:38:32succeeding is low. Very, very low, but
- 1:38:34it is possible. So you have to try to
- 1:38:36find the solution, you have to try to
- 1:38:37fight, meet people who have done it.
- 1:38:39But there you go, it's a good wake-up
- 1:38:41call to go to the US for a week or 10
- 1:38:43days and see people who have ambition;
- 1:38:44it changes a lot of things.
- 1:38:46Thanks, Martin.
- 1:38:47Great wake-up call.
About this transcript
This page contains the full transcript of Guide complet 2026 pour utiliser l’intelligence artificielle dans son entreprise by Le Déclic par Alec Henry, generated from the public captions YouTube serves with the video. The transcript has 22,814 words across 3,321 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.