YT YouTube2Text

YouTube2Text → Transcripts → "Is It Too Late to Get Rich Off AI?"

"Is It Too Late to Get Rich Off AI?" — Transcript

by MoreMozi · 12,722 words · 1,825 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00And so I'm going to break this into
0:01three pieces. All right. AI hygiene be
0:03very quick workflows. Uh and then AI
0:06assets. All right. That's the that's the
0:09that's the deal. All right. So AI
0:12hygiene. Um many of you guys do not
0:16follow AI hygiene and for some reason
0:18you don't know about this. Okay. Um you
0:21want to start a new chat for every new
0:23topic. So, when I first started using AI
0:25a long, long time ago, I basically had
0:27an never-ending one chat that just kept
0:29going forever and ever. And I talked was
0:31like, "Hey, I've got this weird spot on
0:32my arm. What's that?" And then I was
0:34like, "Hey, how do I change this light
0:36bulb in my car?" And then it was, you
0:38know, after that it's like, "By the way,
0:39write me some copy." And like the AI
0:41probably thought I was a psychopath and
0:42it would be correct. Um, but I don't
0:44want it to know that. I want it to infer
0:46that. So for every time that you want to
0:50ask anything that is not related to the
0:52the question before you want to start a
0:54new chat. All right? So just for
0:55everybody number one.
0:58Next the more iterations andor feedback
1:01loops you provide within the same chat
1:04the better the outputs will continue to
1:06be. And so your fourth answer uh on a
1:10string is going to be superior to the
1:12first answer. Again the big C word is
1:14context. All right. Can we get a
1:16favorites chat? I don't even know what
1:17that means, but that that's what we're
1:19looping here. All right. So, if you have
1:21you have your initial prompt that you
1:23added your big context uh to it then
1:25might ask you some more questions. Then
1:27you give that feedback and then it gives
1:29you an output and then you give output
1:30on that feedback. That next thing it
1:32gives will be significantly better than
1:34if you just took the first thing it gave
1:36you. Okay. So, if you want to increase
1:38the quality of the outputs, increase the
1:40number of iterations within each window.
1:42All right. Next,
1:45uh, verify facts and figures of outputs.
1:47We we try to make sure as, you know, we
1:48minimize hallucination as much as
1:50possible, but like you for sure like if
1:52they have if there's numbers in there or
1:53it says a study, make sure that um that
1:56you just double check it.
1:58Um, one important point that I will make
2:01here, and this kind of rough references
2:02Kenan's story, is that like at the end
2:04of the day, decision-m is still yours,
2:07right? You cannot outsource thinking.
2:08You'll still be rewarded by the quality
2:10of your judgment and your taste. So what
2:12we want to really have AI do is
2:13outsource the typing, right? Like if you
2:15know what you want to type and you
2:16already have the ideas, you can just
2:18have someone else clack clack clack
2:19clack on the keys. All right? So
2:21outsource the typing, not the thinking.
2:24Next, the big thing here is that the
2:26idea is still the alpha. So your idea
2:28from a content, from a sales message,
2:30from an offer, the idea is still the
2:32alpha. That is still where the highest
2:34returns come, not necessarily the the
2:37collateral. Now the cloud was a part of
2:38it to describe or or or execute or
2:40implement against the alpha. Um but the
2:43idea is still the core thing and I'll
2:44show you what I mean in a second. Okay.
2:46So that's AI hygiene. Please start new
2:48chats when you have a new idea. The more
2:50iterations you have within a chat, the
2:51higher the the quality of the output.
2:53Verify facts and figures. Don't
2:54outsource your thinking. Only the typing
2:56and the idea is still the thing that
2:57matters most. Okay. Now this gets into a
3:01little bit higher level for those of you
3:03who are who have teams. Okay? And if you
3:05don't have teams, then this will still
3:06apply to you. But just um it's just a
3:09different way of organizing uh
3:10workflows.
3:15Okay, let's rock. So,
3:19so here's a big idea for you. Who here
3:22has heard the term SOP? Standard
3:24operating procedure. Who has heard SOP
3:26before? Just give me a gives me an SOP
3:29in the chat if you heard of SOP.
3:32Okay.
3:35Yes, you guys have heard of SOP. What I
3:38want you to do is in your brain
3:41translate SOP to prompt. It is the same
3:46thing. Think about what it I want to
3:48give you, my employee or myself, our
3:51standard operating procedure for some
3:54task, right? We give them a standard
3:56operating procedure. What do you think a
3:58prompt is?
4:01You're giving them the operating
4:02procedure to get an output. And so the
4:06prompt is the SOP. And so when you think
4:08about it from that perspective, you will
4:10realize that most of y'all's prompts are
4:12too short. You're trying to uh outsource
4:15work to an employee with one sentence.
4:19And so like everything worth doing in
4:21life, it still takes work. Now, I'm
4:24going to talk about the leverage that
4:25this creates later, but big picture, you
4:28still have to do you have to do some of
4:31that thinking work up front to get
4:33higher quality outputs. Like, you still
4:35you can be dumber than you were before
4:37and still be kind of successful, but you
4:39get significantly more returns when
4:41you're smart and work hard. Now, the
4:43rich get richer still.
4:47Now,
4:50when we want to have repeated workflows,
4:52so you have things that occur every
4:54single week within your business. You
4:55make a certain amount of content. You do
4:56a certain amount of outreach. You send a
4:58certain amount of emails. Um, you send
5:00voicemails. You have, you know, RFPs
5:02that you respond to. Um, you have emails
5:05that happen in in and out every day. You
5:07have Slack messages. All of these things
5:09are are that happen on a regular basis,
5:11right? What you want to do is create
5:14standardized prompts for repeated work.
5:17So inside that AI folder that we had, we
5:19had business context which was one
5:21folder. The next folder that we would
5:23have is SOP prompts. Okay. So inside of
5:28those SOP prompts is where we want to
5:30put the prompts for repeated work. Now,
5:34if you're not sure what those are yet,
5:36at the very least, create a list of the
5:39things that you do on a regular basis.
5:41Now, if you don't know how what that is,
5:43then what you'd want to do is a time
5:45study where you can look at what you
5:46spend your time on in a given week and
5:48then block it out at 15 minute slots and
5:50you can see where you spend your time.
5:52Now, the vast majority of that stuff um
5:54is stuff that is repeated. And so, all
5:57we have to do is spend onetime work,
5:59remember I said one time work, forever
6:00payoff or forever time saving onetime
6:02work actually really thinking through
6:04what our SOP would be if we wanted to
6:06hand this off to an employee. What would
6:08it really be? And then that becomes the
6:10basis for the prompt for the AI in the
6:12future.
6:14Yes, ADHD wins again. Okay.
6:18So, you want to store these centrally in
6:19that folder. And if you're the one using
6:22them, um, you can save it into a master
6:24doc with tabs. For those who use Google
6:26tabs, that's one way of doing it. If you
6:27just like love having master docs, I
6:29tend to be a psychopath who likes like
6:30one doc to rule them all. Some of you
6:32guys like folders. That's up to you. I'm
6:33a big like one doc that has a zillion
6:35tabs guy. That's me. Um, but it's up to
6:37you. Now, most of you um like if you
6:40really think about what this would
6:41translate for a role is, you might have
6:42like a customer success manager and
6:44there might be 15 tabs and each of those
6:46tabs relates to one of 15 tasks that
6:48they have to do and the prompt related
6:50to that task. Is this making sense? And
6:53if you're not sure what an SOP looks
6:54like, I will show you I will show you.
6:57Okay? I will show you. Don't worry. We
7:00good so far? Yes. All right.
7:05Now, um, if you are also solo working,
7:08here's an alternative method of
7:09implementing it within your business. If
7:11you have repeated calendar blocks, so if
7:13you're somebody who does time blocking,
7:14which I'm a big advocate of, which is
7:16you block your calendar for when you're
7:17going to do work, um, you can put the
7:20prompt inside of the calendar block so
7:22that when you see it on your calendar,
7:23the prompt is already there and you can
7:25make it a repeated event. All right. You
7:26can also do this for a new employee. You
7:29can time block their time and inside of
7:31their calendar event is the prompt that
7:33they would use to do the work. Pretty
7:36slick, right? Okay. Can I give you an
7:38example? Yes, I'll give you an example
7:40of an SOP in a second. All right. And so
7:42what you want to do is you then can
7:44train the team on using those SOPs,
7:46those prompts rather than just saying,
7:49"Hey guys, use ACQAI, use the LLM,
7:53right? Go use them." It's like that's
7:54not helpful for somebody. Just like
7:56saying, "Go make emails." They need more
7:59context so they can make them the way
8:01you want them to be made. And to be
8:03clear, this is just like uh this is just
8:05using AI like any other machine with a
8:08checklist.
8:15And so the end result here of this
8:18second second little portion here is
8:20that you have a prompt repository. You
8:23have an area where your SOPs are saved
8:26and these become even more powerful new
8:28age SOPs. So think of prompts as just
8:322.0 0 SOPs 2.0, right? And so your
8:35business functionally is run by a series
8:37of prompts. There is work that must be
8:40done. There is a way to do that work.
8:41And we save the prompts for how to do
8:44that work properly.
8:48Okay?
8:49If you do it many times, then it means
8:52you should only do it once and do it
8:54right and then replicate it. I'm going
8:56to say that again. If you do the same
8:58thing over and over again, then you
8:59should do it one time, right? so that
9:01you don't have to do it again over and
9:03over again.
9:05Brian, how do we know when to update the
9:07prompts? When you stop getting outputs
9:09that are good.
9:12Okay, now this may be the the meiest
9:16part of today. Okay, and I said I was
9:18going to keep today shorter, and that's
9:20because this is already a shitload of
9:21work. All right, so I like my goal here
9:23is just like I want you to actually do
9:24this. Okay,
9:27this is the bigger one. All right, so
9:30big picture, we got our context window
9:31101 and 2011. That's what we did first.
9:34Next, we make sure we follow our AI
9:36hygiene. We start new chats. We iterate
9:38more times within the chat in order to
9:39get better outputs. We're good on that.
9:42Then for the repeated tasks that we do
9:43within the business, we start creating a
9:46prompt repository SOP list. And you can
9:49do this by roll. And you can do it on
9:51their calendar, you can do it within a
9:53Google doc, or you can do both. So
9:55people know exactly what steps to follow
9:57in order to do the work.
10:00Now the third piece here is AI assets.
10:04All right. This is this is going to be
10:07probably the besides the context window.
10:10This will be the highest ROI. So I
10:11probably put the SOP thing I just said
10:13as as third highest ROI. This would be
10:15the second highest ROI thing that we
10:16will do today. Okay.
10:20So we already have the folder of AI
10:23assets. So we have our business our
10:25business context number one. We have our
10:27SOP or prompt repository number two. And
10:30then number three we have AI assets.
10:32Sorry, we have um I mean you could call
10:34them I guess it's AI assets within AI
10:36assets but um
10:38I guess probably have to call a
10:39different name. Um you call it data
10:41sources if you want to. Um but
10:45call it data sources for for right now.
10:48Sorry I I I messed that up by saying
10:49calling the same name twice. Um data
10:52sources. All right. That's what the
10:53folder will be. Now
10:56what you want to do is go by function
10:58and gather the assets. Now what does
10:59that mean? Okay. So what we do is we
11:02will export the things that you have
11:04already done. So the big advantage that
11:06you have especially the longer you've
11:07been in business is that you have done
11:10stuff. You have worked and that work has
11:13created output. That output is data.
11:16That data can be used to create more
11:18data like it in the future.
11:20So
11:22you should have a place where you have a
11:24source that has all of the data of the
11:26work that you've already done so that it
11:28can help you create more work that looks
11:30like that.
11:33So this is a sample list of assets that
11:35I would start with
11:37if those of you who tweet. I would put
11:39all of your tweets which you can export
11:41in one button on X. You export every
11:43single one of your tweets and that
11:45becomes boom your prop. that becomes,
11:47excuse me, a data source. Number one, I
11:50put all my tweets all time data source.
11:52Real quick, I'm going to show you the
11:54exact 10-stage road map from zero to 100
11:57million plus that less than 1% of
11:59companies finish. I've now done multiple
12:01times. And so, I can say with a lot of
12:02confidence that these are the stages as
12:04headcount increases that you need to get
12:06through. And I broke each of these down
12:08by eight different functions of the
12:10business. What the constraint feels
12:11like, like, what are the symptoms of it
12:13when you're going through it, and then
12:14what steps we actually took to graduate.
12:16And we've done this across software,
12:18physical products, uh, service
12:20businesses, brick-andmortar, all of
12:22this, and it works. And it's my gift to
12:24you. It's absolutely free. And so, the
12:25links in the description, just enter
12:26your info and it'll spit it right back
12:28to you. All free. Number two, if you've
12:31sent any emails to your list over the
12:33last however many years of your of your
12:35customers, take those emails and then
12:38put all of them in one document, one
12:39master doc that just has the emails that
12:42you've ever sent. Now, alternatively, if
12:44for each of these that I'm giving you,
12:46if you have certain emails that you
12:48like, you can use those emails,
12:51right? If you like an e-commerce store
12:53that sends a certain type of email or or
12:55whatever, you could just search all
12:57those, then export, and then boom, you
12:59have another data source.
13:03By the way, the calendar prompt is not a
13:04prompt. It's just you paste in the
13:07prompt and you just put it in the
13:08description of the calendar event. um so
13:10that when that person clicks the time
13:12they can see the prompt, they can import
13:13the prompt, they can do it, they can use
13:14it. So it's not like a special thing.
13:16I'm just saying it's a different way of
13:18organizing it.
13:20Another asset that you should probably
13:22have for those of you who have sales
13:23teams, you probably have SDRs or
13:25setters, right? You probably have those
13:27calls recorded. You already have this
13:29stuff. Take your top 10, top hundred if
13:32you want recordings, the transcriptions
13:34of those calls, and then we want to make
13:37that another data source. Now, what I'm
13:39putting in the bold next to this as
13:41we're going through this is what you can
13:42create from these assets. Right? So, if
13:45I have my top 10 sets of all time
13:48transcribed, I'm going to give you the
13:49prompt momentarily that you can use to
13:51transform that into the golden standard
13:53of your SDR script. If you have your uh
13:57email newsletters that you've done in
13:58the past, that helps you create your
14:00future newsletters. If you have your all
14:02your expost and tweets that you've done
14:03in the past, that helps you create
14:05future expost and tweets.
14:08If you have your top 10 closing calls
14:10recorded, this will help you create your
14:12VSSLs. It'll create your emails, it'll
14:14create your follow-up, create your
14:15closer script. So, what do I use for
14:17extracting uh transcriptions if it's in
14:19a different language than English? I
14:20mean, I don't think it matters the all
14:22the LLM speak the same language. They
14:23they're universal. They're they're large
14:25language. Well, you can speak in
14:26whatever language you want.
14:28Um,
14:30if you have uh lead nurture uh
14:32sequences, uh you can export all of
14:34those things and then say, "Hey, this is
14:36my current lead nurture. How could I
14:38improve this? How could I make this
14:39better? Output superior versions of the
14:41lead nurture." So, I'm going to keep
14:43going through these because I know that
14:44like SDR means setter, the person who
14:46qualifies leads. Um, I'm going to keep
14:48going through these, but the the big
14:50takeaway that I want everyone to have
14:51from this is that you already have these
14:55data sources. you have already done this
14:57work and so it's so the advantage that
15:01all of us who have done work in the past
15:02have for using AI is that you can get
15:06the AI to sound like you because you
15:09have done stuff. So I'll uh I'll pause
15:12real quick. Who here in the chat has
15:14seen content on the internet and been
15:16like this was written with AI? Just give
15:19me a give me a give me an eyeballs
15:22eyeballs emoji. If you've seen AI
15:26content on the internet just like words.
15:28Just the words. Not even the videos.
15:29Just the words. Yes.
15:32Of course. Now the question is why?
15:38Why is it AI content? How can we tell
15:41that it looks like AI content?
15:45The reason we can tell it's AI content
15:47is because AI sounds like a a melting
15:51pot of the internet.
15:53And so AI sounds like the internet. And
15:58so if you want So AI doesn't sound like
16:00AI. AI sounds like us
16:02if we had no personality.
16:05And so if you want the AI to have a
16:07tone, have a a pattern of speaking, have
16:10a certain cadence that it that it
16:12communicates in, the advantage that you
16:15have, all of us who have actually done
16:17work, is that you can use the work you
16:19did and say, "Don't sound like you,
16:21sound like this."
16:24And that it will use that and pattern
16:26match to create the output that you
16:29want. And so AI doesn't like if people
16:31like because I hear this a lot like like
16:33AI can only make stuff that sounds like
16:34AI. No, AI sounds like AI if you give it
16:37no context. You have given no asset, no
16:39data source to work off of. Then it's
16:41just going to fabricate from the large
16:43data set that it has which is the
16:44internet. And so that's why these
16:46filters, that's why we start with these
16:48context uh up front so that it can know
16:51what pattern to match with the outputs.
16:56But wait, there's more. Thank you, Lee.
16:58Okay. And so for each of these that I'm
17:00showing you here, it's you've got the
17:02data source and then we've got the thing
17:04that we can create from the data source.
17:05And I'll give you the pro I'll show you
17:06the prompt stuff in a second. Okay? But
17:08each of these, this is the relationship
17:09I want you to see is that like if we
17:10have if we have sales recordings,
17:12there's lots of stuff we can do with
17:13that. We can improve a gold standard
17:15script. We can make a video sales letter
17:17taking the best closes we have. We can
17:19take those sales scripts, put them into
17:20emails. We can do it make it into
17:22follow-up. all of these first ass these
17:25assets or these data sources can create
17:27further assets and data sources in the
17:29future. Okay. Um if you want to help if
17:32you if you're like hey I want to make
17:33better content use transcripts from
17:35YouTube videos that you like
17:39and yes uh Himeme or Jamie um I'll show
17:42you how to do the the tone blueprint in
17:44a like we'll do that at the end.
17:47Uh transcripts from YouTube videos that
17:49you like. Great. We take those
17:51transcripts and then we feed it in and
17:53that gives us more videos like this. All
17:55right? And I'll show you again the more
17:56tactical ones, but this is I just want
17:57to give you conceptually the idea here
17:59so it can get your get your juices
18:00flowing because every single one of you
18:02right now, you guys tweet, you write
18:05emails, you have you have sales people,
18:07you have VSSLs, video sales letters, you
18:09have emails, you have follow-up, you
18:10make content, you make all this stuff,
18:13right? And if you're not using this to
18:16make that, then you're not you're not
18:18using the ACQA to the at all the same
18:21degree that you could be and getting way
18:22more out of it. Okay. Um product FAQs,
18:26right? That transforms into sales page
18:28copy, VSSL, FOB, lead, long-term
18:30nurture, uh customer support
18:32transcripts. So, when we're thinking
18:33about this folder that we're talking
18:34about, we've got our assets, we click
18:36in, we've got our context, right? We
18:38have our SOPs and prompts for the
18:40repeated work we do in the business. And
18:41then we have the data sources that those
18:44SOPs and prompts are based off of. Are
18:46seeing how this all put together.
18:49If I can
18:52is there an overhead cam that's on? If
18:54there isn't, then about it. I can use my
18:56hands.
18:57>> There's not. Okay. Okay. So then pay
19:01close attention to my hands.
19:05We're going to double click on this
19:07first folder, which is your AI assets.
19:10Inside you will have three folders.
19:12Folder number one is your business
19:14context. That is all the data that
19:16you're going to have on your business.
19:17Folder number two is you're going to
19:19have your prompt repository, your SOP
19:22repository. These are the prompts that
19:24dictate how you do the work you do on a
19:26regular basis in your business. These
19:28are instructions.
19:30The third folder is going to be the data
19:32sources. This is what these prompts will
19:36reference to pattern match to create
19:38outputs.
19:41All right.
19:43Does everyone get all three of those?
19:47Everyone everyone with me so far? Yes.
19:52Business context.
19:56SAP repository
19:58data sources.
20:00That's it. It's all we're doing. Your
20:02SOPs are the things you do and how you
20:03do them. Business context is what you've
20:06already done and the description of your
20:07business.
20:09And your data sources are where you are
20:13gathering, you're exporting data from
20:15things you've already done. Your tweets,
20:16your emails, your STR recordings, your
20:18sales recordings, your existing email
20:19sequences, uh video transcripts,
20:21transcripts of videos you like, emails
20:22you like, right? Content you enjoy,
20:25product FAQs, reviews online, customer
20:27support transcripts, all of that stuff.
20:29All of those are data sources.
20:34your best performing ad copy
20:38to give you a few others. If you've got
20:40high engaging uh social media posts that
20:42you like or have seen from other people,
20:44keep a one, you know, one doc that has
20:47the the post that you like the most.
20:48That rhymed. Uh your website copy in and
20:52of itself is an asset, right? It's a
20:55data source that it can pull off of,
20:57which can also give you page
20:58improvements, copy improvements, ad copy
21:00from there, emails from there. If you
21:02have blog posts or articles that you
21:04enjoy, those become data sources. SOPs
21:07and internal processes. Some of you guys
21:08already have some SOPs. Great. Now, I
21:11will I don't I'm going to hit that memo
21:13system. I I think I might I might blow
21:16someone's gasket. So, I might not do
21:18that one right now. I think it's going
21:19to it's going to melt some minds. Okay.
21:21We'll just move on to the next one. Um
21:23onboarding scripts for customers or
21:24employees, right? These are all things
21:26that you have.
21:29pitch decks and webinars, slide decks
21:32that you have, all of these new data
21:35sources, top performing email sequences,
21:39guys. Are you guys with guys with me on
21:41this? You want that that repository to
21:46be rich with the highest quality data
21:48you can. And the higher the quality that
21:50you curate with those sources, the
21:54higher quality the AI will pattern match
21:56to create outputs. And the more complete
21:59the context, the more specific and
22:01personalized they'll be to your
22:03business. And the more detailed the
22:04prompt, the more they will follow the
22:06instructions, give you the output you
22:07want. That's how all three of these
22:09things work. All right? It has to know
22:11what your business is about. It has to
22:14have good data that it can base off
22:16saying, "This is what it looks like when
22:17it's right." And it has to get clear
22:19instructions on exactly what you want.
22:20When you have all three of these things,
22:22you can put them together and make
22:24banger outputs. And so things only sound
22:27like AI when you don't have a baseline
22:29to train it on. So AI sounds like the
22:31internet. But if you train it on the
22:32work you've done, it sounds like the
22:33work you've done.
22:37Yes, I will give you guys examples. I
22:39want to give you I want to walk through
22:40this first. Okay. And so these are the
22:43most common things that people are using
22:46uh to get outputs with ACQAI. So people
22:49are looking for sales scripts. So right
22:51now many of you I know because I've
22:54talked to you guys many of you need one
22:56or more items on this list right so who
22:59here um actually what do we have we got
23:0212 items on the list who here say on
23:05number one to 12 how many of these
23:07things you know you could do better at
23:11than you are right now
23:13you could do better on one of them do
23:15better on two of them five of them all
23:1712 put a number 1 through 12 of these
23:19things are on this Sales scripts 12. 12.
23:23We got a lot of twelves. Okay. That was
23:25not actually where I was going for that,
23:26but that's good to know that a lot of
23:2712. Three, right? A lot of twelves.
23:29Five. A lot of twelves.
23:32Um, some fours, some seven, six. Okay.
23:35Use a lot. This is brilliant. Okay,
23:37cool. So, the reason I bring this up is
23:40that this this now gets to the crux of
23:43the issue of like what separates ACQAI
23:45from kind of the generic AI, right?
23:47which is that for many of you, these 12
23:51things are the majority of how you make
23:53money, right? You make content, you run
23:56ads, you do outreach with words and
24:00videos and media, and then those people
24:03become engaged leads, and then you give
24:05them more information via sales scripts
24:08and video sales letters and webinars.
24:10And then with that extra information,
24:12they then make an informed decision and
24:14then purchase your stuff. And then after
24:16that they get more information from you
24:19in some way. Right?
24:22And so it's not that do you have this or
24:25not have it. It's how much do you have
24:27and how good is it? And this is why the
24:29difference between a 225 batter and a
24:31300 batter is such an important point.
24:35So the whole idea here is that we're
24:36trying to make the asset HQI that can
24:39help improve the quality of these things
24:41as they come out based on all the things
24:43that we've trained it on. Now this is
24:44clearly just on assets. There's
24:46obviously, you know, offer strategy,
24:48there's decision-m, there's things like
24:49that that it also is trained on. But for
24:51many of you, when I look at the top use
24:52cases that you guys have, these are the
24:55these are the big use cases you guys use
24:57it for. All right. So, if many of you
25:00guys did say at least one of those
25:01things is something that you could
25:02improve on. Okay. The reason this is
25:05important, why this matters is that if
25:07more better is how you grow, this allows
25:10you to do that. You no longer have an
25:12excuse because you now have tremendous
25:15leverage. And I just found this little
25:16infographic that I thought was funny.
25:17All right, boss, leader, introvert,
25:19introvert with AI, right? I just thought
25:22it was cute. And so some of you guys
25:24asked, how do I use this? All right, how
25:27do we put out so much stuff and how do
25:30we keep the quality high? Here's how we
25:32do this. So Mosy Minute used to take me
25:35between 60 90 minutes per email. And so
25:39this generates, put this in context,
25:41this generates about 20% of the revenue
25:43of ACQ, uh, the advisory practice. All
25:45right.
25:47Now, it takes four hours for me to make
25:5016 emails. So, basically, one half day
25:52per month, I sit down and I write all
25:54the emails for the month. And so, I get
25:57So, this gives me 4x leverage
26:00using what I just outlined. I've got my
26:02business context. Again, what are the
26:04three pieces here? I probably should
26:05have put a slide up, so I apologize.
26:07There's three big pieces. We have
26:09business context. We have the rules of
26:12engagement. We have how we do the stuff.
26:14And then we have the stuff we base it
26:15on, our data sources. Those are the
26:18three things that you need to have to
26:19use AI. Well, and so using all three of
26:22those things, I wrote the Mosy Minute
26:24for a very long time before using AI.
26:26And so then I said, great. And I'll tell
26:29you, it's like AI does not come up with
26:30the idea though. This is a big one. AI
26:33does not come up with the idea. It will
26:35take the idea that I have and say, "Hey,
26:37um, I want to write, uh, uh, a Mosy
26:40Minute based on this core offer
26:43structure. This is why it's cool. This
26:44is why I think it's awesome. Turn this
26:46into an email that looks like these ones
26:48and follows that structure and then it
26:49gives me the output and then I tweak and
26:51edit it. Thank you. I appreciate that
26:53that I still spend time on it. It's just
26:55the difference is that I get more time
26:57back." So many of you guys think in
26:59terms of automation when you really
27:01should just be thinking in terms of
27:03leverage and amplification. I want to
27:05say this again because I think it's a
27:06really important point.
27:09Many people want to sell you an
27:11automation and it's and hey it's sexy as
27:13hell. Automation's sexy but more
27:16realistically what you will get is
27:17amplification not automation. You will
27:19get more leverage not infinite leverage.
27:22True automation is infinite leverage
27:24right? It's it's it just it continues to
27:27run. It never has to get checked. It
27:28never breaks. It always improves. That
27:31is uncommon. What is far more common is
27:34that something that used to take me 60
27:35to 90 minutes now takes me 15 to 20
27:38minutes.
27:40Real. So, I'm getting 4x. Thank you,
27:42Kevin. I appreciate that. Um, and to be
27:44fair, Kevin, it's not AI. It's me. It's
27:48my idea. It just helps me type the words
27:50faster um and give examples faster,
27:53right? So instead of me sitting there
27:54being like, "All right, what example do
27:55I want to use here?" It just gives the
27:56example. I'm like, "Great. It already it
27:57already used the idea that I had it and
27:59applies it, right?" And so they're not
28:01AI, I take the emails that I have
28:04already written that write that I write
28:06like me. I have the idea that's me. I
28:09say, "Go make this similar to that. It
28:11gives me the output and then I still
28:12edit it and tweak it, right? And the
28:14tweets, for example, that start the
28:15email with, those are tweets that I pull
28:17from somewhere else, right? And so all
28:19of this is leverage. All right. Next,
28:23um, tweets themselves used to take me
28:25about an hour and a half a day. And to
28:27give you context on this, it gets 20 to
28:2930% of all of our content comes from
28:31tweets. All right? And so now it takes
28:33me about 90 minutes a week. So
28:35basically, I'm seven times more
28:36efficient. I haven't automated this.
28:37It's not AI making my tweets. Is that it
28:40will help me generate ideas from which
28:42to iterate to make tweets that are good,
28:44right? So if I could just purely say,
28:47"Hey, make me all these tweets." And I
28:48copy pasted them, it it wouldn't work. I
28:50like believe me I feel like it doesn't
28:51work right but it can give me
28:53inspiration uh and ideation to base off
28:56of to then make my stuff happen faster
28:58because before this I just sit there and
29:00be like all right and I would scroll I
29:02would scroll X I'd scroll uh social
29:04media and usually I would just wait
29:06until I felt an emotion about something
29:07and be like why do I feel this way and
29:09then I would respond to that content as
29:12my tweet right
29:15and technically Stephen no it's not
29:17still written by AI because I'm still
29:18still editing a lot of it um it would be
29:20assisted. Um, if it's pure AI, it's
29:23different. So, which is a a specific
29:26difference by the way from a from a
29:27copyright perspective. Um, so next is um
29:33YouTube, right? So, videos used to take
29:35two hours for video for me and this is
29:37like where 40% of lead flow comes from,
29:39right? So, now it takes 15 minutes per
29:41video. And so, what's interesting is
29:42that many of you will take the word AI
29:44and think, oh, bad quality. That's
29:45because you don't know how to use it.
29:46It's also because you don't have
29:47existing resources. This is a big point.
29:50This is a big deal.
29:52You are rewarded for the work that you
29:54have done.
29:56My minute emails are bangers because I
29:58wrote a shitload of bangers. My tweets
30:01are good because I already wrote 6,000
30:03[ __ ] tweets before I did this. 6,000.
30:06All right, I'm at 7,000 now. 6,000 was
30:09just me, right? And so it has a huge
30:12datab data source to pull from to give
30:15inspiration in terms of sentence
30:16structure in terms of ideation that
30:18allows me to go work a lot faster. And
30:20so I gave my estimated Yes. I I do a lot
30:23of work, Dylan.
30:26Yes, Emil. Exactly. I put 6,000 in and
30:28said, "Learn from this right
30:32now." Um,
30:35if we're going through here, notice not
30:37all of the amplification is the same,
30:39right? Right? I got 4x leverage on
30:41emails. I got 8x on YouTube.
30:45Different, right? And the key point
30:47here, big big point. It's the idea. The
30:51idea is the leverage. The ultimate alpha
30:54is the idea. And so, it's not going to
30:57come up with the mossy minute because it
30:59can't have my experience. And so, I know
31:02some of the questions that I'm going to
31:03get around this are like, is AI going to
31:05replace us?
31:07eventually not for a while but the main
31:11reason is that because AI cannot live in
31:13real life and cannot get real life
31:15experiences and so those experiences are
31:17the alpha that's the idea that's the
31:19learnings those learnings those stories
31:22those narratives it can then flesh out
31:24but the core idea is yours that is the
31:27essence of it and so this is how this is
31:30how I've scaled what I do is like I can
31:33take the core the core value and then I
31:37can amplify It it's not automated. It's
31:39just amplified.
31:41You guys digging this? Is this making
31:43sense?
31:48Correct. Moritz A is not AI is not
31:51original yet. It does. It's like a
31:52remixer, right? Exactly. For now, it
31:56will get better in the future, but for
31:57now, it it more or less remixes. And so,
32:00the idea is what I want to Okay, I can
32:03now articulate this well.
32:05What we're trying to remix is our
32:08language patterns, our sentence
32:09structure, the format of the video. I'm
32:12gonna have an intro, I'm gonna have
32:13three main points, I'm gonna have
32:14stories, whatever. Right? The format,
32:16the structure, that is what AI is very
32:19good at replicating
32:21the essence of it, the core value that's
32:24being created. That sounds like very
32:25amorphous language for me. Um,
32:26fundamentally, the good [ __ ] right,
32:29that's inside that still is going to
32:31come from you. Which is why a beginner
32:33who just tries to churn out AI content
32:35still doesn't win because they have no
32:38idea. They have no experience. They
32:40haven't done anything. And the AI can't
32:43do it for you. Can't live your life for
32:45you. Can't have the experience for you.
32:47So, you have to do stuff like I've
32:48always said and then take the stuff you
32:50did and you talk about it. AI just
32:52allows you to talk about more of it
32:54faster.
32:56Yes, Rayhan.
32:59We can amplify. That's the goal here.
33:02All right. Um, memos used to take me at
33:04least four hours per memo. Um, now they
33:07take me probably an hourish,
33:11right? Um, maybe a little less than
33:13that. I don't know. Maybe maybe was
33:15maybe it was a little aggressive on
33:16that. Probably still takes me about an
33:17hour. Um, so it was 4x lever 8x
33:20scripts for sales guys. My god, that
33:22would take me a long time. Um, and now
33:24we can write scripts much faster for new
33:26offers, for portfolio companies, for
33:28anything, right? is we get significantly
33:30more leverage.
33:32And so I'm bringing this up for you guys
33:33to show you that like how do I use AI to
33:36make YouTube videos. It's going to be
33:38more structure and ideiation. Um and
33:41then honestly I have a skill set which
33:43is that I've developed over many years
33:44which is that I can riff on topics in a
33:46way that's interesting and I know syntax
33:48and rhetoric that people find engaging
33:51being real. Um so my stuff isn't
33:54scripted and that's what also makes it a
33:55lot easier. Um, so for me it's much more
33:57like what are the what are what are the
33:59interesting prompts that drive response
34:01around a topic and that is easier. So
34:03like I still so the way I use AI is
34:06actually to prompt me to say interesting
34:09things. Actually that's probably the
34:10easiest way for me to say this. Like the
34:12tweets for me is that it prompts me to
34:15respond to things which helps me
34:17ideulate on interesting things to say
34:19faster. It does not say, "Hey, Alex, say
34:21this."
34:24Now, for for scripts, yes, you can use
34:27it to create sales scripts. All right?
34:28So, that's what we're going to get into
34:29next. All right? And so, yes, VSSLs used
34:31to take a full day, and now you can do
34:33them really quick. Okay?
34:36So, the part that many of you guys have
34:38been waiting for. So, big picture, we've
34:40we've broken this down into number one,
34:43we have our business context. Number
34:45two, we have our SOP repository of how
34:49we do [ __ ] And I'm going to give you
34:50examples of this in a second. And number
34:52three, we have our data sources, right?
34:55Which is stuff that we've done in the
34:56past that it can reference to make more
34:58stuff like it. Okay. What if you're not
35:01confident about the quality of your
35:02current content? Excellent question,
35:04Casper. I will give you use cases for
35:06how to improve this.
35:09Yeah, AI uses me the way other people
35:11use AI. Yes, I get AI to prompt me.
35:13Okay, so the goal here is we use past
35:16work to make future work. So this is the
35:17overarching step process. So number one
35:20is you get assets, you separate them
35:22clearly. So what I mean by that, so
35:23let's say uh let's say you got 10 sales
35:25call transcripts. You just got to make
35:26sure that each sales call is clearly
35:28delineated like space space star begin
35:32script star begin third script star star
35:35begin fourth script. All right as you go
35:38through. So you get your assets, you
35:41separate them clearly and then you
35:42upload them. All right. So this is the
35:43data sources we're talking about. So you
35:45upload that, you put your little paper
35:47clip, right? You upload it. Then you
35:50prompt, right? The prompt repository
35:52that I was referencing. I probably
35:53should have done step three instead step
35:55two first, but you get the idea. Okay.
35:58So you'll notice that the best outputs
36:01come from the best inputs. And when done
36:02right, these create prompt chains. And
36:04I'll explain what this means later. So,
36:07for example, if you have a sales
36:09recording but don't have a gold standard
36:12sales script, let me um let me send
36:16something. Hold on. Liz, can you pull
36:18something up um on
36:21on the screen? Can we go like back and
36:23forth between two screens? Is that
36:25possible?
36:31>> I I'm send you a doc because I want to
36:33show what the prompt looks like.
36:35um so that they can they can see it. So,
36:39just so you guys know, I'm going to give
36:40you Yeah. I mean, so believe me, this is
36:43like the one-on-one training, so I'm not
36:44going to do any more than this. I don't
36:46I don't want you to be overwhelmed.
36:47Again, before anyone gets too deep in
36:49this levels here, okay? Number one is
36:53get your context window right. Do 101
36:56and 2011. Get that first. If you if you
36:58do nothing else but doing that, it's a
37:00W. The second thing you can do is just
37:03get your data sources. Everyone has
37:06them. There's there's no work that's
37:07really involved there besides you just
37:09capturing it. Just put it together. All
37:11right, that's the second big thing. The
37:13third thing is the prompt repository
37:14that references both of these things to
37:16create outputs.
37:19Thomas, there is a isn't there a limit
37:21providing context on every one of the
37:22platforms? There is, but this is why I'm
37:24saying try do things in word format and
37:27that'll dramatically decrease the size.
37:29If you're talking it's 25 megabytes. If
37:31you're doing just words, you'll have
37:32more than enough. You could put a book
37:33as context with 25. Again, the key thing
37:35is you don't put images inside. Okay.
37:37So, I'm going to have my team um uh go
37:41through this. So, I had this originally
37:44as in the slide deck, but I think I'll
37:46just voice over it. I think it'll be
37:47more useful for everybody. Um yes, thank
37:50you, Buck. We've worked on it. Um okay,
37:53so let's say right now, um you have some
37:57sales recordings. Okay, you have some
38:00sales recordings, um, but you don't have
38:02a unified script. So, your sales team
38:04doesn't have a unified script. So, what
38:06we would do is we would paste the
38:08transcript of the 10 call recordings and
38:11then we would post this prompt inside of
38:13it, right? And so, it' be like, hey,
38:16this is the number one. This is for you
38:17guys. And I will give you guys a copy,
38:19my gift to you, of every one of these
38:21prompts, uh, as an email. Okay? So, you
38:23guys get an email with every one of
38:24these. Um, but I want to go through each
38:26of them so you understand how this works
38:27because fundamentally you've got stuff.
38:30We give it a prompt to make better stuff
38:32or analyze that stuff and improve it.
38:34That's it. All right. So, you're an
38:36elite B2B sales trainer and script
38:38doctor. I will paste 10 full transcripts
38:40of successful sales calls that sold the
38:42same offer to the same type of prospect.
38:44Your job is to build one mother script
38:46from these calls by extracting only what
38:48actually moves the sale forward and
38:49deleting everything else. Follow these
38:51rules. So, you can scroll up. So analyze
38:54these calls first. Identify common
38:56structure but best calls follow from
38:57greeting to close. Note the exact
38:59questions and phrases that consistently
39:00leads to open disclosure uh commitment
39:02and payment. Define the call structure.
39:04Build the script in clear sections with
39:06headings. Opening rapport frame and
39:07agenda discovery. Transition to pitch
39:09offer and value stack. Handling three to
39:11five objections. Final close and next
39:12steps. All right. Booking. Bam. Bam. All
39:15right. Create the unified script. Use
39:18only lines that appear in multiple
39:19winning sales calls or are clearly
39:20essential glue. Remove filler, small
39:22talk, or stories that don't directly
39:24help the sale. Keep language single.
39:26Keep language simple, direct, and
39:28conversational. Third to sixth grade
39:29reading level. Uh, keep the core pitch
39:31portion short enough to be said in under
39:32five minutes at a normal pace. Where
39:34useful, include questions the rep should
39:35ask verbatim and notes about intent or
39:37tone. Uh, output format, a short
39:39overview of the structure in bullet
39:41form. Uh, final word for word script
39:43organized by section headings. Bullet
39:44list of the top 10 most powerful phrases
39:46or lines you pulled from the calls.
39:47Remember, prioritize clarity, brevity,
39:49and closing power. Delete everything
39:50that doesn't help prospect decide and
39:51buy. All right. So, this is what you can
39:54take your transcripts and turn it into
39:57sales scripts. Transcripts. Sales
40:00scripts. Transcripts. Sales scripts.
40:05This is cool, right? This is pretty
40:08sweet, isn't it?
40:11Yes. Dope. Dope [ __ ] Okay, cool. So, I
40:16said I was going to give you guys
40:17examples. like we're going to this is
40:19this is the meat. All right, this is the
40:20meat. All right, I know transcript sales
40:23are pretty sexy. All right, so now let's
40:25say you uh want to write more emails. So
40:28right now, you know you should be
40:29following up, right, with your leads.
40:32You know you should be following up with
40:33your leads. You know you should have a
40:34newsletter because Alex said he's
40:35getting 20% of his revenue just from
40:37spending one half day a month. And
40:38you're like, man, if I spend one half
40:39day a month and I and I got an extra 20%
40:41revenue, uh that would be make me money.
40:43And I would say I agree with that. Uh
40:45and you should do that. So, how do you
40:47do it? All right. Uh, this is a
40:50newsletter style and uh and so basically
40:52you're going to analyze newsletter and
40:54generate. So, it's both. So, you're an
40:56expert copyriter, narrative strategist.
40:57You'll notice the same structure for
40:58Lilies. You give them the role, you give
41:00the task, you give rules, and you give
41:01output. It's kind of the four the four
41:04kind of structure for good prompting.
41:05All right. So, you're an expert
41:07copyriter, narrative strategist, train
41:09to analyze writing samples, extract
41:10voice, structure, style patterns, and
41:12reproduce them with new ideas while
41:14maintaining the same feel and reader
41:15impact. I'll provide my past email
41:17newsletters that performed well. Your
41:19job is to learn the voice, tone, pacing,
41:20narrative structure, rhetorical
41:21patterns, and call to action style from
41:23these examples. Analyze my past so
41:25tasks. Analyze my past winning
41:27newsletters. For all newsletters, I give
41:28you to extract voice profile, writing
41:30ticks, narrative structure, format, and
41:31patterns, persuasion mechanism, CTA
41:33style. Provide concise but complete
41:35descriptions of the style you infer.
41:38Next,
41:39build a reusable newsletter template
41:42using the patterns above. Create a
41:43generalized structure for my winning
41:44newsletter style, including opening hook
41:46style, story and body pattern, teaching
41:47framework, pacing and rhythm rules,
41:49closing style. This becomes the
41:51newsletter blueprint. Three, generate a
41:53brand new newsletter after analysis when
41:55I give a new topic idea. So, this is
41:57this is the alpha, right? Remember guys,
41:58I was like, you still have to come with
42:00the idea. So, you say, "Hey, I want to
42:02write." So, uh, I want to write a
42:04newsletter on some stuff that roofers,
42:08you know, so people who own roofs, uh,
42:10or people who have leaky pipes,
42:11whatever, that they can, uh, do avoid
42:14having leaky pipes in the winter. Easy.
42:16So, I want to have this specific thing,
42:18this core story or core element that
42:20happened. And I'm going to say, hey,
42:22that's the story. And I say, great.
42:23Boom. Put in my template. Go run it.
42:26Right? And then it's going to fill in
42:27the gaps uh, within that structure. is
42:29going to take that story and amplify it
42:31so that it actually finishes follows the
42:33structure that I want. So then the
42:34output will be one full newsletter match
42:36the same style use the template um and
42:38then it should feel like I wrote it and
42:40so then it has rules. So, some of you
42:42guys never copy phrasing from old
42:44newsletters. Learn the style, not the
42:45words. Newsletter must be 100% original
42:47content in my voice, right? That's why
42:48the key idea is so important. If you
42:50actually just ask it to just write
42:53newsletters for you guys, it will suck.
42:55I'm just telling you, it will suck. It
42:57does not work. If you give it the idea,
43:00then you will have a valuable a valuable
43:02insight. Keep theru structure
43:04consistent. Never break character. Um,
43:06and if you need more samples to refine
43:08the model, ask. Yes, Simon, I'm going to
43:10give you all these my gift to you. All
43:12right. U and then this is output format.
43:15So you can just scroll past this.
43:17Okay. So here's another one. So let's
43:20say you have notice it's like if you
43:23have this then you can do this, right?
43:25So uh let's say you've got a uh
43:28um you're you're getting you're getting
43:30stuck on rebuttals, right? You're like
43:32people are are are not buying from you
43:34and they're coming up with objections,
43:35right? So we say, "Great. Let's help
43:38solve this. We're going to upload our
43:40transcripts, right? And now we're just
43:42going to focus on the objections. So,
43:44you're an expert sales and objection
43:45handling strategist. You specialize in
43:47reviewing failed sales calls,
43:49identifying the core objection patterns,
43:50and turning them into clear, repeatable
43:52objection handling playbooks. The input:
43:54I'll provide you with one or more failed
43:56sales transcripts. Each transcript may
43:57include hesitation, fill words, messy
43:59phrasing, or incomplete responses. Your
44:00job is to extract and clean the
44:02insights. Task: Identify the primary
44:05objection each call. So, each script I
44:07give you give you the primary objection.
44:08Secondary objection. Quote the exact
44:10lines from the transcript that signal
44:11the objection. Summarize the objection
44:13in one clean sentence. Two, propose
44:15strong objection handling sentences. So
44:17for each objection, write three to five
44:18short, powerful objection handling lines
44:20the salesperson could say next time.
44:22These must be conversational,
44:24non-confrontational, confidence
44:25building, framed around value, clarity,
44:27and reducing perceived risk, including
44:29both softening statements and direct
44:31reframes. Three, build a consolidated
44:34objection guide after processing
44:35multiple calls across all transcripts.
44:38you analyze. Create a running playbook.
44:40Identify all objections that appear more
44:42than once. Cluster them in the top 10
44:44most common objections and for each of
44:45the 10 give a clean definition of the
44:48objection. List example phrases
44:49customers say and provide three to five
44:51objection handling lines. Provide the
44:52root cause behind the objection. Fear,
44:53confusion, misalignment, timing, money.
44:55And then format the final guide in clean
44:56way that can be copied into your
44:58training manual.
45:01Right. Okay. There's more output. I'm
45:05not going to read the output formats for
45:06you guys, but that's because then it you
45:07just tell it how you want it to, you
45:09know, deliver the thing. So you you
45:10scroll the next one. Okay. Um, so next
45:14one. One of you guys said earlier, I
45:16wish there was like some sort of uh
45:18writing style analyzer, right? This is
45:21what it looks like. So if you've written
45:23stuff anywhere via Slack, via email,
45:26blog posts, uh any any any content you
45:28put out, you can deduce your writing
45:32style and then you can feel you can you
45:35can put inputs in so it doesn't sound
45:36like AI, it'll sound like you. All
45:38right, so role is you're a world-class
45:40writing style analyst. Your job is to
45:42dissect the style of a given text so it
45:43can be reverse engineered or reused.
45:45Focus on how it's written, not what it
45:46says. I'll paste simple writing after
45:48this prompt. Analyze it deeply and
45:49systematically. Number one, high level
45:51snapshot. So overall style summary,
45:53author persona, and then primary goals
45:55of the writing. Next, give me the voice
45:58and tone, formal, informal, direct or
45:59indirect, conversational or
46:00authoritative. Uh tone shifts. Does it
46:03stay consistent or does it change? Any
46:05notable moments where a tone becomes
46:06intense, emotional, humorous, serious?
46:08Relationship to the reader. Is this
46:09because it's their peers, expert, coach,
46:11storyteller, teacher, or something else?
46:12Is the voice empathetic, challenging,
46:14supportive, or confrontational?
46:15Structure and flow. Number three, how is
46:17the text built? All right. So, overall
46:20uh structure pattern. So, is it hook
46:22story, lesson framework, CTA? Are there
46:23distinct sections or acts? Is there
46:25opening? Is there closing? Is there
46:26logical flows? Right? How we work
46:28through this stuff? Sentence uh sentence
46:31level style. So, drill into how
46:32sentences are written. So, sentence
46:33length and rhythm, syntax and
46:35construction, punctuation habits, use of
46:37questions, right? Diction and word
46:39choice. We scroll through this. Do Yeah.
46:43Is there repetition? Are there
46:44analogies? Are there metaphors? Are
46:45there rhetorical devices? Notice how
46:47much detail there is on this one. We're
46:48trying to get the we're trying to get
46:51the true essence of the writing. Most of
46:54you guys are just saying sound like me.
46:56We have to be specific. That's the idea.
46:58It is a it is a machine, right? So, you
47:01give it machine-like instructions.
47:03Imagine you gave this to a human. The
47:06the what's interesting about this is
47:07that a human would have a very hard time
47:09maintaining all this context in their
47:10mind when they're trying to write. A
47:12robot can. And that's what makes this
47:14powerful.
47:16Cool. Okay. Keep show keep for me. So
47:20style and rules the checklist. You can
47:22keep going
47:24and then emulation sample and then we
47:26have the output format again. All right.
47:28And so then this becomes our next thing
47:31um that you can complete like completely
47:34plug in to anything else.
47:37Yes, this is talking to an autistic spe
47:40specialist, right? [laughter]
47:43Um real talk. Everything I do I take
47:45from that perspective more um when I
47:47give instructions that's the whole
47:49point.
47:51Um you can use these uh prompts
47:53obviously in HQAI which is what I'm
47:55talking about. But fundamentally so like
47:57I'm like I have I want everyone to win.
48:00That's my whole goal here. Um I use all
48:03the LLMs including including ACQAI. I
48:07use ACQAI the most for my own business
48:09strategy questions because it pulls from
48:11the most context of business a business
48:14right rather than like this is pulling
48:15from blogs on Wikipedia and stuff like I
48:18want stuff that is the most applicable
48:20for a small business which you know for
48:22our purposes will be you know zero to
48:23100 million um and so especially
48:27strategy questions especially like offer
48:29based questions pricing questions sales
48:31flow sales process it's exceptional at
48:34very good with sales process uh Don't
48:36sell GPT sound like me again. Yes,
48:38please don't. Okay, so with each of
48:40these other ones, you can get the idea
48:41for how much depth there is here. I'm
48:43going to just read the um the headlines.
48:45Um so that you can get an idea for how
48:47this works. So um for some of you guys,
48:50you guys have sales transcripts and
48:52recordings, which if you don't, please
48:54do that first, but you don't have VSSLs.
48:56You can take your sales transcripts and
48:58say, "Hey, you're a top top tier direct
48:59response writer and sales strategist.
49:01You specialize in converting long-term
49:03sales transcriptions into polished,
49:04high-converting VSSLs.
49:07So, you upload the document containing
49:0810 successful sales calls and your job
49:10is to extract persuasion patterns,
49:12emotional beats, objection overcomes,
49:14promises made and transform uh and
49:16transformations described and convert
49:17them into a cohesive cinematic VSSL
49:19script. And so, many of you guys need to
49:22just take this, take your 10 sales
49:23calls, use this to transform them into a
49:26VSSL script, and then read the damn
49:28script. All right, let's go to the next
49:29one. Just scroll all the way to the next
49:30one.
49:34Okay. So, uh, number six. So, uh, sales
49:38call manager, sales call feedback
49:39analyzer. So, this is a big one. So,
49:41many of you guys have sales teams and
49:43you're too lazy to give sales, uh, sales
49:46feedback. So, you can have the team one
49:48of two things. One is the sales manager,
49:50if that's you, you take their
49:51transcript, feed it in, and say, "Hey,
49:53give feedback as like give me the
49:55feedback on this sales script as though
49:57you're a master salesperson."
49:59So then when you go into your
50:00one-on-one, you can say, "Hey, this is
50:01what you need to work on." And then you
50:02can roleplay it. Now, if you need to do
50:04that for the whole team, you can do
50:05that, too. Guess what else you can do?
50:07You can have the team get their own
50:08feedback on their own sales call. Or you
50:11can take your own sales script and get
50:12your own feedback from a master
50:13salesperson,
50:17right? Okay. Let's go to the next one.
50:24Okay. So, if you like specific style of
50:28YouTube videos, right? You're like, I
50:30like this style. I like this structure.
50:32Um, take YouTube videos you like, and
50:36paste it in. And then you would include
50:38what you want it to be about. And then
50:40it would pattern match to make a video
50:41that's about a new topic with your idea
50:43with that general structure and tone.
50:46Right? So, you're a world-class YouTube
50:48script analyst and creator. You study
50:50video transcripts to extract structure,
50:52tone, pacing, narrative mechanics,
50:53emotional beats, comedic timing, if any,
50:55hooks, story arc, CTA patterns, and
50:57signatures or sty stylistic choices. I
50:59will upload one to five transcripts from
51:01YouTube that I like. Your job is to
51:02analyze them, extract the style,
51:03structure, tone rules, build a reusable
51:06script template for when I give you a
51:08new idea, generate a brand new script
51:10using that template, matching the exact
51:12feel of the original videos, but with
51:13completely new content.
51:16Dallas, how do you get transcripts from
51:18YouTube? You literally there's a button
51:20on videos on YouTube to get them. Like
51:22literally it's literally a button. So in
51:24the description you hit see more and if
51:28you scroll there'll be a button that
51:29says see transcript and then you can
51:31just copy it. It's like you don't have
51:33to have a tool. You can just like
51:34YouTube gives you the transcripts. So
51:37you copy two or three of those and make
51:38sure again with when you whenever you
51:40have multiples make sure that you
51:41delineate star star you know number one
51:44star star star number two star star
51:45number three as you keep going. Okay,
51:48great. Do I prefer HQA for writing copy?
51:51Yeah, I mean I use it. Um I'm biased
51:55though. So, um yes. Uh I like it for
51:59writing copy. Again, I like I'll be honi
52:04the most around offer, pricing, sales
52:07flow, copy. It's probably the big four
52:09that I use it the most for. Um
52:12but anyways, let's keep going. So, uh,
52:15this would then output, um, YouTube
52:17stuff from past YouTubes that were good.
52:20So, keep going. Okay. Uh, shorts
52:23scripts. So, for those you that's that's
52:25the longs format. This is if you have
52:27shorts format. So, either take the your
52:29best performing shorts or take shorts
52:31that you like that you like the style of
52:33and then you insert those into our
52:35[snorts] data source and it already has
52:37the context window of your business. And
52:39so, then you say, great. Boom. Here's
52:41the prompt.
52:44You're a world-class short form content
52:45analyst and creator. You break down
52:46YouTube shorts, Tik Toks, reels, and
52:48understand their hook style, pacing,
52:48emotional, beat, jump, cut, rhythm,
52:50punch line, and timing. Creator voice
52:51upload one to I will upload one to four
52:53uh one to five short form transcripts,
52:55videos I like, you analyze the
52:56transcripts, exact structure, and tone
52:58and pacing rules. Build a reusable short
52:59script template. And when I give a new
53:01idea, generate a brand new short form
53:03script using that style. Final script
53:05must feel indistinguishable and vi from
53:07the original content while containing
53:09completely new content.
53:12All right, let's go to the next one.
53:18I like Gemini for images. For those of
53:20you curious, I like Gemini and GPT for
53:21images. Um,
53:25I like
53:28I use those I use a lot for images. I
53:30use GPT a ton for like personal
53:32questions. So, more like health stuff.
53:35Like if I like I'll give you a real
53:37story. My uh microwave kept getting this
53:39child lock on it. So, I took a picture
53:41of it and I said, "Can I how can I
53:43remove this?" And it was like, "Oh, this
53:44is series E17722.
53:47Here's the owner's manual. It says this
53:49is how to remove it." I was like,
53:50"Great." Stuff like that. I use it for
53:52those types of things. Um, I use Grock
53:55for anything that I want to not have any
53:58kind of political agenda woven into it.
54:01So, anything that I have that I want to
54:03ask Grock about is like more like edgy.
54:07[laughter] If I have like hypothetical
54:09questions that I don't need the GPD to
54:10judge me, u I'll ask rock. Um, so yeah,
54:14that's that's like what I use those ones
54:15for. All right, back to this. So, uh,
54:18many of you guys, uh, need lead magnets.
54:21Okay, so many of you guys have videos,
54:23have content that you've made in the
54:25past. You take those transcripts, uh,
54:27you take those SOPs, you take those
54:29assets that you already have, and then
54:31you plug it in to create a lead magnet
54:33for your business.
54:36So you're world-class content strategist
54:37and direct response marketer. You
54:38specialize in transforming video content
54:40into high converting lead magnets
54:41designed to create demand, build
54:42authority and generate mount leads. I
54:44will upload one to three videos or
54:45transcripts of content I like. It may be
54:47educational, motivational tactic or
54:48storybased. You will take this tone,
54:51structure, frameworks, insights and
54:52teaching patterns and then build a
54:53master lead magnet based on those
54:54patterns but with new original content
54:56aligned with my topic and audience.
55:01Yes, Morates. Exactly. If you want to be
55:02the sole emperor of the world, how would
55:04I go about this? Exactly. Okay. So,
55:06that's if you have this stuff and need
55:08lead magnets. Let's go to the next one.
55:12This is the last This is the second to
55:14last main one and then I'll show you the
55:15last kind of advanced thing. And I think
55:16we're I think that probably I'll have
55:18enough minds melted for the day. Um so,
55:22uh once you have the lead magnet, like
55:25once you give a mouse a cookie, it's
55:26going to ask for some milk, right? So,
55:28once you have the lead magnet, what do
55:29you need next? You need the follow-up
55:31for that lead magnet. And so you say,
55:32"Hey, you're an elite email and SMS
55:34conversion copywriter. You specialize in
55:35turning lead magnet downloads into book
55:37sales calls using this big stack. I will
55:40upload my lead magnet PDF or text. Your
55:43job is to read it, extract the
55:44messaging, and create a compelling or
55:46sorry, complete follow-up system that
55:47converts readers into book calls."
55:50Right?
55:54That's the idea. So, some of you guys
55:55have lead mags. You're like, "It doesn't
55:57work. No one books calls." It's like,
55:58well, have you asked it to make an
56:00idealized um uh follow-up sequence?
56:04Right? That's SMS and email that's
56:07proven for that specific lead magnet.
56:12That's what this helps you do. Okay,
56:13let's go to the next one.
56:17Okay. Uh,
56:20keep going. Keep going. Okay. Offer
56:21architect. So, this is one of the big
56:23ones for you guys. Um, this is one of
56:25the like based on the use cases that
56:26we've already pulled from the AI, this
56:28is one of the biggest ones that you guys
56:29use at the most commonly. Um,
56:33so this is going to be a little bit
56:35different. So 11 through 14, the ones
56:37I'm going to show you are actually a
56:39prompt chain. All right? So what's the
56:41difference here? So the first 10 I
56:43showed you were more or less like if you
56:45have this, you know, basically use this
56:47prompt plus your context plus the data
56:49source to get a good output. All right,
56:50that's the idea.
56:5211 through 14 basically stacks or loops
56:55the output to create the input of the
56:57next one. Okay, so output one becomes
57:00input two in output two becomes input
57:04three output three becomes input four.
57:06All right, this is the best succeed. I'm
57:08glad Daniel glad to hear that. Okay, so
57:11this is the offer architect. So some of
57:12you guys like need to sell more stuff or
57:14people aren't buying as much as you
57:15want. And so this is what we'd start
57:17with. So you're an offer architect,
57:19world-class strategist. You think like
57:21pricing economist, behavioral
57:22psychologist, blah blah blah blah. Your
57:24job is to absorb any offer I upload and
57:26turn in more powerful, more
57:27irresistible, higher converting, higher
57:28profit, and more defensible offer. So,
57:31I'm going to read this one in total
57:32because I think it's important and then
57:33we'll go to the other ones. All right.
57:34So, step one, analyze my current offer.
57:36When I upload my offer, extract core
57:38promise, mechanism, value drivers,
57:39tangible outcomes, intangible outcomes,
57:41time to result, effort required, risk,
57:43pricing structure, guarantees, bonuses,
57:44delivery model, avatar, sophistication
57:46level, how it's currently positioned
57:47versus the market. Then create a
57:50diagnostic report showing strengths,
57:52weaknesses, missing elements,
57:54marketability, risks.
57:58That's step one. Step two, now build a
58:01more irresistible version of my offer.
58:03So, offer enhancements, add or improve,
58:05guarantee bonuses, fast action
58:07incentives, delivery upgrades, reduce
58:08reduce steps or complexity, increase
58:10time to value, uh, increasing perceived
58:12certainty, increasing tangibility, uh,
58:14social proof or evidence, insert
58:15recommendations, value stack, right?
58:18I'll go through all I'm just going to
58:20skip that one. Okay. Pricing
58:21architecture. Here's optimal pricing.
58:23Here's our anchor pricing, middle entry,
58:24bundles, hybrid pricing, payment plans,
58:26etc. Step three, add upsells, downells,
58:28or cross sales. This is optional. Boom,
58:30boom, boom, boom, boom.
58:32Step four, build the grand slam offer
58:34version. Create a completely new
58:36rearchitect version of my offer that is
58:38clear, simple, dramatically blah blah
58:39blah blah. Keep going through. Um,
58:42include the title, one sentence,
58:43promise, target avatar, primary
58:45mechanism, delivery model, what they
58:46get, line by line, bonuses, guarantees,
58:47pricing, reasoning behind the pricing.
58:49That's a key one by the way. Uh, and CTA
58:51framing. Step five, high converting
58:53offer positioning, messaging. So now
58:55give the elevator pitch, headlines, risk
58:57reversal statement mechanisms. Keep
58:59going. Step six. Now we have the output
59:02format and we can skip the output. So is
59:04each step a new prompt? No, this is one
59:06prompt. This is the thing is like you
59:08guys are putting like two sentence
59:09prompts in and somehow expecting it to
59:11write a book for you. Like let me give
59:13you let me give the let me give the way
59:15it works. The longer the prompt, the
59:18shorter the output. The shorter the
59:20prompt, the longer the output.
59:24It's kind of like the shorter the
59:25question, longer the answer. Like what's
59:26the meaning of life? Very short
59:27question, very long answer. If I said if
59:29you are a you know uh you know old white
59:34American man born in this time married
59:37to this person here's a book about that
59:38person what is that person's meaning of
59:40life it might be a sentence
59:43the idea is the more context you give
59:44the more it narrows the answer so many
59:46of you guys probably let me let me know
59:47in the chat actually if you guys get
59:49annoyed when you put a simple question
59:52in and you get like two pages of answers
59:54you're like I just needed [ __ ] yes or
59:55no Right.
59:58Do I ever get tired of cranking so much
1:00:00value? Brian, it's a great question. Um,
1:00:03I don't know. I'll tell you. We'll we'll
1:00:05find out by the end of the day. Um, yes,
1:00:07you guys get that. I get super annoyed,
1:00:09too. Why do you think that is? Short
1:00:11prompt, long response. Long prompt,
1:00:14short response.
1:00:17That's what it is. You have to give more
1:00:19context so that it answers exactly what
1:00:21you want. It doesn't know what you want,
1:00:23so it just gives you a ton of answers,
1:00:24hoping that one of them matches.
1:00:26Okay, now remember what we just did
1:00:30number 11 is that we got the new offer,
1:00:31right? The output was the offer. Yes,
1:00:34I'm tired of it now. Thanks, baby. So,
1:00:36we got the So, remember output, input,
1:00:38output, input. So, the the output of
1:00:41prompt 11 was the prompt of number 12.
1:00:45So, we now have our offer. What do we
1:00:48need? We need our sales page copy
1:00:50generator. So, we say, okay, I have my
1:00:53offer. Now, I got to write the copy. So
1:00:54your world class copyriter specializes
1:00:56in writing high converting sales page
1:00:57based on a clear structured offer. So I
1:00:59will paste my offer architect output
1:01:02including all of these things. So we
1:01:04take the output and we paste it in as
1:01:06the input and then scroll. We can just
1:01:09scroll all the way. That's fine. And
1:01:10then it's going to output all these port
1:01:11portions. Now the output of that one is
1:01:14going to be our sales page. And so we
1:01:15say hey with the output of our offer
1:01:18architect I want you to also create my
1:01:20sales call script. Right? Because we
1:01:22have our architect. You gave us the
1:01:23output. We say, "Great. You're a
1:01:24world-class sales trainer and call flow
1:01:26architect. You turn offers into high
1:01:28converting sales call scripts. I will
1:01:29paste my offer architect output. Your
1:01:31job is to create complete sales call
1:01:32script that a rep can follow." Boom. Now
1:01:34we have our sales call script off of
1:01:36number 11.
1:01:38Yes. The chain prompt. Yes. And then
1:01:41number 14. Great. Now that we have our
1:01:42offer and that we have our sales page
1:01:44and we have our call script, what else
1:01:46we need? We need our VSSL. So we take
1:01:48our offer architect. We put the output
1:01:50in and then boom, we have our VSSL. And
1:01:52so just like that, when you can chain
1:01:55these together, you can build out entire
1:01:58marketing universes the way you want it.
1:02:01And the key point of all of this stuff,
1:02:03guys, is the three wheels I started this
1:02:06whole day with, which is we have to have
1:02:09the 47 questions of context around your
1:02:11business, right? And so you say, now
1:02:12I'll paste my offer, create the VSSL
1:02:14using the structure buff, right? So
1:02:16whenever you see those things at the
1:02:18bottom, that's where you clearly
1:02:19separate. That's where star star start
1:02:20star and then you put the put the inputs
1:02:23right. Um [snorts]
1:02:24so all of these will work if you have
1:02:27wheel number one complete context for
1:02:29the business wheel number two great
1:02:31props which is what we're saying clear
1:02:33instructions to follow and then rule
1:02:34number three the data source that it can
1:02:37pull from to make good quality outputs.
1:02:40Those are the three wheels of making AI
1:02:42work significantly better. People can
1:02:44tell the difference between truly
1:02:45automated work, which is why nothing I
1:02:47have sounds like I mean doesn't nothing
1:02:49I have sounds like AI, right? I just use
1:02:52it because it helps you get more.
1:02:58That's the game, right? The game is more
1:03:00now. That's the game. If you can do 10
1:03:03times the work, which is the part that
1:03:04you guys I think many of you guys are
1:03:05missing is like every one of you should
1:03:07be able to 10x your output with this and
1:03:09still keep your quality the same. That's
1:03:11the key. If you like I know a bunch of
1:03:13you guys are like I love Moy Minute. Me
1:03:14too. I spend time on it. I just don't
1:03:16spend 90 minutes like I used to. I spend
1:03:17like 20. Now if some of you are like
1:03:20well he only spends 20 minutes on it
1:03:21now. Well I don't know. I do Q&A things
1:03:23that are five minutes and people pay a
1:03:24lot of money to do stuff like that
1:03:26right? And so like
1:03:29I just don't have to waste my time on
1:03:32lower value work. So if I know what the
1:03:34core idea that I want to talk about is
1:03:37once the core idea is there flushing out
1:03:40some of the words and the sentences and
1:03:41transitions and like bolding the stuff
1:03:43like that stuff is a pain in the ass. It
1:03:44takes time. It's not valuable right the
1:03:47value is in the idea the values in the
1:03:49narrative the values in the story that's
1:03:51the value in the experience and that's
1:03:53what everyone who's watching this you
1:03:54guys you have the experience you guys
1:03:56are doing stuff in the real world you
1:03:58need to document that. That's what the
1:03:59data sources are for. We have to have
1:04:01these data sources so that you can have
1:04:02something that's unique, right? We want
1:04:04to use the AI like a workhorse, right?
1:04:07Use it like a workhorse, not to make
1:04:09your decisions for you. Real quick, I'm
1:04:11going to show you the exact 10-stage
1:04:12road map from zero to 100 million plus
1:04:15that less than 1% of companies finish.
1:04:17I've now done multiple times and so I
1:04:19can say with a lot of confidence that
1:04:20these are the stages as headcount
1:04:22increases that you need to get through.
1:04:24And I broke each of these down by eight
1:04:26different functions of the business.
1:04:28what the constraint feels like, like
1:04:29what are the symptoms of it when you're
1:04:31going through it, and then what steps we
1:04:32actually took to graduate. And we've
1:04:34done this across software, physical
1:04:36products, uh, service businesses, brick
1:04:38and mortar, all of this, and it works.
1:04:40And it's my gift to you. It's absolutely
1:04:42free. And so the link's in the
1:04:43description. Just enter your info and
1:04:44it'll spit it right back to Offering.

About this transcript

This page contains the full transcript of "Is It Too Late to Get Rich Off AI?" by MoreMozi, generated from the public captions YouTube serves with the video. The transcript has 12,722 words across 1,825 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 translate the transcript into 100+ languages.

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, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.