YT YouTube2Text

YouTube2Text → Transcripts → I’ve Built 500 AI Workflows, This is What Businesses Want in 2026

I’ve Built 500 AI Workflows, This is What Businesses Want in 2026 — Transcript

by Nate Herk | AI Automation · 3,812 words · 575 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00After building hundreds of AI workflows
0:01for real clients, and teaching thousands
0:03to do the same, I noticed something
0:04weird, which is that most people online
0:06are building the fancy stuff, but
0:07businesses don't actually want that.
0:08They just want five types of
0:09automations, and these are simple,
0:10boring workflows. They're things that
0:12genuinely save time, save money, or
0:14remove mistakes. And these five
0:15automations are the ones that clients
0:16will pay the most for. So, in this
0:18video, I'll show you what they are, why
0:19they sell, and how you can build and
0:20sell them, too. But before we jump into
0:22the first one, let me show you how I
0:23figured all this out. I've worked with
0:24coaches, real estate agents, dentists,
0:26HVAC, e-commerce, and more. And I've
0:28seen the same patterns across every
0:30business. They're different industries,
0:31yes, but they have the same problems.
0:32And these workflows keep showing up
0:33again and again. So, if you're watching
0:35this, and you want to make money selling
0:36AI in 2026, don't overthink it. Just
0:38start with what businesses are actually
0:39paying for, what the data is showing.
0:41So, let's get into the first workflow.
0:42So, this one is speed delete. And
0:44honestly, it might be the single easiest
0:45automation to sell to any service-based
0:47business. So, here's what it is. When
0:49someone fills out a form on a website,
0:50sends an inquiry, or calls a business,
0:52there's a clock that immediately starts
0:53ticking. And the faster that the
0:54business responds, the more likely they
0:56are to close that deal. Studies show if
0:57you respond to a lead within 5 minutes,
0:59you're up to 10 times more likely to
1:00convert them, compared to responding
1:02even 30 minutes later. And yet, the
1:04average business takes 47 hours to
1:06respond to a new lead, which is two full
1:07days. By that point, the prospect has
1:09already talked to three of your
1:10competitors, or gotten they reached out
1:11to you, or maybe they've just solved the
1:13problem already themselves. So, a speed
1:14delete system fixes that gap completely.
1:16The moment someone submits a form, the
1:17automation kicks in, it captures the
1:19lead info, it qualifies them based on
1:21criteria like budget, location, service
1:23type, whatever, and then it routes that
1:24lead to the right person on the team.
1:26And it also fires off a personalized
1:27follow-up message immediately. So, all
1:29of this happens within seconds. So, who
1:30actually pays for this? Service-based
1:32businesses. Dental clinics, law firms,
1:33home services companies like HVAC and
1:35plumbing, real estate agents, even
1:37marketing agencies. Basically, any
1:38business where a missed lead or a slow
1:40response directly equals lost revenue.
1:42So, basically, every business. And the
1:44reason they pay so well for this is
1:45because the ROI is immediately obvious.
1:47You don't have to convince them of some
1:49theoretical AI agent future benefit that
1:51may or may not happen. You just show
1:52them the math. So, let me give you a
1:53simple example. Let's say you're working
1:54with a local dental clinic. They're
1:56spending five grand a month on Google
1:58Ads to drive new patient inquiries. And
2:00let's say they're getting 100 new leads
2:01a month from those ads. But because the
2:02front desk is busy, they're not getting
2:04back to people for hours, sometimes not
2:05even until the next day. Their close
2:07rate on those leads is 12%, which means
2:09out of 100 leads that come in, they're
2:10booking 12 of those as new patients.
2:12Now, you can build a speed delete system
2:14where every time someone fills out that
2:15appointment request form, the system
2:16instantly sends a personalized text and
2:18email, which confirms the inquiry,
2:19qualifies them based on the type of
2:21service they need, and it notifies the
2:22right person at the clinic right away
2:24with all the context that they need. So,
2:25what you did there is you didn't change
2:26the ad spend, you didn't change the
2:28copy, you didn't change the offer, you
2:29just made the actual business faster.
2:31And let's say that close rate goes from
2:3312% to 25%, which is conservative. So,
2:35they're getting 13 extra patients a
2:36month for the exact same amount of ad
2:38spend. And the key is, when you can walk
2:40a business owner through that exact math
2:42using their data, price objections will
2:43disappear. Because at that point, if
2:45they really think about it, the
2:46expensive decision is to ignore you and
2:48just not hire you. All right. So, now
2:49that you understand that first one, the
2:51second one's going to make more sense.
2:52And this is probably the least exciting
2:53automation on the list, but it's also
2:55one of the most profitable. And that is
2:56document processing. So, first, let me
2:58just paint you a picture. Imagine you're
2:59an accountant, and it's Monday morning,
3:00so you've got 200 invoices sitting on
3:02your inbox from last week, and your job
3:04for the next several hours is to open
3:05each one, read it, pull out the vendor
3:07name, the amount, the date, the line
3:08items, and then just type all of that
3:09into a different software, categorize
3:11it, and then file it, 200 times. Every
3:13single week, you have to do that. And
3:14that's not an exaggeration, that's the
3:15actual reality for a small accounting
3:17firm that I've worked with. 50 hours a
3:18week, basically a full-time employee's
3:20entire job, just moving numbers from
3:21PDFs into spreadsheets. So, at $30 an
3:24hour, let's say, that's $78,000 a year.
3:26So, what you can do is build a workflow.
3:28Invoice systems can come in via email,
3:29the system extracts the vendor, the
3:31amount, the date, all these other things
3:32that you need. It can also check it
3:34against a chart of accounts, it can flag
3:36anything unusual, and then push all that
3:37clean data where it needs to go. This
3:38will cut processing times down from 15
3:40minutes per invoice, let's just say, to
3:42something like 2 minutes, because you
3:43still maybe want like a human review
3:44step at the end. And that would free up
3:46about 45 hours a week, which is over
3:48$70,000 in direct labor savings per
3:50year, before you even factor in things
3:52like human errors, because those cost
3:54the business money, as well. And the
3:55cool thing about document processing
3:56automations is that a lot of times, you
3:58don't even need AI for this. Some of the
4:00most valuable document workflows that
4:01I've ever built were purely rule-based.
4:03No LLMs, no fancy prompts, just clean
4:05logic that moves data from A to B
4:07without a human ever touching it. And
4:08when there's no AI involved, the system
4:10is completely deterministic, it runs the
4:11same way every time, it's rock-solid
4:13logic, and it's basically
4:14maintenance-free. And the reason this
4:15sells is simple. Manual document
4:16processing takes about 15 minutes per
4:18document, costs between 15 to 25 dollars
4:20each, and has an error rate between 5
4:22and 15%. And like I said earlier, every
4:24one of those errors actually costs real
4:25money to the business. So, you automate
4:27that, you free up some time, you drop
4:29the cost, and the math does the selling
4:30for you. And unlike speed delete, which
4:32is mostly service-based businesses, this
4:34one hits a completely different set of
4:35industries. Insurance companies, law
4:37firms, accounting firms, logistics,
4:38construction, really anyone that's
4:40drowning in paperwork. If people in the
4:41business are spending hours every day
4:43just moving information between systems,
4:45then there's a deal waiting to be made.
4:46All right. So, before I get into this
4:47next one, I need you to understand
4:49something. I'm about to cover two
4:50workflows back-to-back that sound pretty
4:52similar on the surface, which are
4:53follow-up sequences and database
4:54reactivation, but they solve different
4:56problems. Follow-up is about leads who
4:58just came in and need more touchpoints
5:00to actually convert. And reactivation is
5:02about leads that you forgot existed, you
5:03know, 6 months ago. They're two
5:04different situations, they're two
5:05different systems, but they are both
5:07extremely profitable. So, pay attention
5:08to the distinction, because it does
5:10matter. All right. So, follow-up and
5:11nurture sequences first. This ties
5:13directly into speed delete, because
5:14getting the lead fast is only half the
5:16battle. What happens after that first
5:17touchpoint is where businesses also,
5:19once again, drop the ball. Because most
5:20businesses spend a ton of money
5:22acquiring leads. They get someone
5:23interested, maybe they fill out a form,
5:24or book a call, or attend a webinar, or
5:26something like that. And then, the
5:27business follows up with them once,
5:28maybe twice, and then they move on.
5:29Meanwhile, the prospect just needed one
5:31more touchpoint before they were ready
5:32to actually buy something. Research
5:33shows that 80% of sales require at least
5:35five follow-ups, but most sales people
5:37just stop after one or two. And here's
5:38the thing that kills me. That person has
5:40already raised their hand. They've
5:41already said, "Hey, Mr. or Ms. Business,
5:43I am interested in what you're doing."
5:44And that first intent action is always
5:46the toughest part of the entire sales
5:47process. Because once someone has filled
5:49out a form, or attended a webinar, or
5:51had a conversation, they're now a warm
5:52lead. And converting a warm lead is
5:54exponentially easier and cheaper to
5:55convert than trying to convince a cold
5:57stranger to care. And yet, most
5:59businesses just let those warm leads go
6:01cold again, because they don't follow
6:02up. So, a follow-up automation can fix
6:03this. When a trigger event happens, the
6:05system will kick off a personalized
6:06sequence automatically, every single
6:08time. And over time, you can layer on
6:10more intelligence, you can pull in CRM
6:11data, you can do real-time research,
6:13whatever it is, but you can add more
6:14steps into this nurture sequence. And
6:16honestly, I would kind of fit this into
6:17the bucket of appointment setting. So,
6:19think like an Instagram DM automation
6:21that will nurture people who followed
6:22your account and try to book them into a
6:24call with you. Now, I'm not going to run
6:25through a full industry list again. I
6:27think you already understand the types
6:28of businesses [music] that need this,
6:29you know, coaches, consultants,
6:30agencies, stuff like that. But the sweet
6:32spot specifically is a business that
6:34already have decent lead volume, and
6:35they know that they're not staying in
6:36front of people well enough. Because
6:38obviously, you can't follow up with
6:39leads or reactivate leads if their
6:41bottleneck in the first place is getting
6:42leads. But here's a simple example. So,
6:43let's say a B2B consulting firm runs
6:46monthly webinars. 150 registrants, 60
6:48show up. Before automation, someone is
6:49manually sending follow-up emails and
6:51trying to book those calls. And this
6:52could take two to three days, because
6:54the messaging is inconsistent, half the
6:55list never gets touched, and humans just
6:57forget. manual process had a conversion
6:59rate of, let's just say, 4% of
7:01registrants. [music]
7:02But after the automation is in place,
7:03every attendee is getting a personalized
7:05follow-up within minutes. No-shows get a
7:07different message with the replay of the
7:08webinar. Over the [music] next 2 weeks,
7:10the system runs three to five
7:11touchpoints with real value, each
7:13pulling in context, so it always feels
7:14personal. The second someone replies or
7:16books a call, the sequence stops, and
7:17the sales team gets notified with all of
7:19the context of that conversation. The
7:21conversion rate will jump to maybe 10 to
7:2212%, and that's 18 calls instead of the
7:25original six that we were getting with
7:26the manual process. And then, of course,
7:28you turn that into a value. So, if the
7:29average deal is $20,000, and they close
7:3130%, that's going from $36,000 to over
7:34$90,000 per webinar. Same webinar, same
7:36content, same spend, they just actually
7:38followed up. Okay, so that was follow-up
7:40sequences, which handle new leads that
7:41need nurturing. Database reactivation is
7:43a completely different animal. This is
7:45about going back in time. Every business
7:46that's been around for any amount of
7:47time is probably sitting on a goldmine
7:49that they've completely forgotten about.
7:50Past customers who have churned,
7:52newsletter subscribers who never bought,
7:53people who started a free trial and then
7:54just disappeared, leads who went quiet
7:56after a conversation. Same concept as
7:58before, these people already know the
7:59business, they've already had some level
8:00of interest, and they're just sitting in
8:02a CRM collecting dust. So, this type of
8:03automation, database reactivation, will
8:05pull them from that existing database,
8:07it will segment people based on where
8:08they dropped off, and can send
8:10personalized outreach. This is designed
8:11to restart the conversation, or
8:13reactivate the lead. When someone
8:14responds with interest, the system
8:16qualifies them, hands them off to a
8:17salesperson as a warm lead. No mass
8:20blasts, no generic messaging, each
8:21touchpoint references their specific
8:23history with the business. Now, let me
8:24skip straight to why this is so
8:25powerful, because I think an example
8:27will hit harder than me explaining the
8:28mechanics. So, let's take a local gym as
8:30an example. Let's say they've been open
8:31for 3 years. They've got 4,000 people in
8:33their system, which is past members,
8:34trial users, and cold inquiries, but
8:36nobody's reaching out to any of those
8:38people. They're spending all of their
8:39budget on new ads, trying to get fresh
8:41people in the door. So, you build a
8:42reactivation workflow that segments the
8:44list and sends personalized messages
8:45referencing each person's history with
8:47the gym. And even a modest 2 to 3%
8:49conversion rate here means that that's
8:5180 to 130 people walking back through
8:53the door. At $50 a month with an average
8:558-month retention, that's 32 to 48,000
8:57dollars in recovered revenue. With no
8:59new ad spend, no new copy, no new leads,
9:02just working with what they already had.
9:03And agencies that specialize
9:04specifically in database reactivation
9:06are reporting average ROI of 1,200% in
9:09the first 60 days. So, this one almost
9:10sells itself. You ask a business owner
9:12how many contacts are sitting in their
9:13CRM doing nothing, you show them what
9:15even a small percentage of those people
9:16coming back would mean for their bottom
9:18line, and you're not selling an
9:19automation anymore, you're showing them
9:20basically just a big pile of money that
9:22they forgot they had, and you're
9:23offering to go pick it up for them. The
9:24industries that benefit the most from
9:26these are ones with high lifetime value
9:28and recurring revenue. So, gyms, dental
9:29clinics, SaaS platforms, e-commerce,
9:32coaching, anyone with a database of 500
9:34or more contacts that aren't being
9:35actively worked. All right. So, the last
9:37one, and I saved this one for the end,
9:39because it might be one of the most
9:41underrated automations on the entire
9:42[music] list, which is internal
9:43reporting and status notifications. It
9:45doesn't sound too sexy at all, but it's
9:47one of the stickiest automations that
9:48you can sell. Because once a business
9:49has it and is using it, they literally
9:51could not imagine going back to life
9:52without it. Same way that you and I
9:53would feel if we lost our favorite AI
9:55productivity tools, we would feel super
9:57slow. Every business has people spending
9:58hours compiling information that the
10:00other people need to make decisions. So,
10:02sales managers pulling weekly pipeline
10:03numbers, agency owners compiling client
10:05KPI reports, operations people gathering
10:07status updates from all these different
10:08tools. None of this work is hard, once
10:10again, but it's just manual and it's
10:11repetitive and it eats up hours every
10:13single week. And I can tell you
10:14firsthand, when I was working at Goldman
10:15Sachs, a massive part of my job was
10:17automating exactly these types of
10:19reports. And that's one of the reasons
10:20why I got into the automation space
10:21because I realized how much time can
10:23truly be saved. So, an internal
10:24reporting automation can handle stuff
10:26like [music] this in the background
10:26because it can pull data from different
10:28systems, it can run the analysis, and it
10:29can just basically notify the team where
10:31they would have gotten that already.
10:33This could be a daily Slack message with
10:34yesterday's sales numbers, it could be a
10:36weekly email with the client performance
10:37metrics, it could be an automatic
10:38notification when a deal hits a certain
10:40stage or if a project falls behind
10:41schedule. The system runs on its own, it
10:43pulls real-time data, and it puts it
10:44exactly where the team already looks for
10:46information. So, no new dashboards, no
10:48new tools, no new processes. You're just
10:50automating a process that sometimes
10:52people have their entire full-time job
10:54just doing that. And once again, I think
10:55that it's not to replace that person,
10:56but maybe let that person do something
10:58that's a little bit more valuable to the
10:59business. So, here's an actual example
11:00of one of the most profitable
11:01automations that I've ever seen, and it
11:03was so, so simple. It just converted
11:05daily phone orders into the same text
11:07format that the construction crew was
11:08already using. It got organized
11:09automatically instead of someone typing
11:11it every morning. It saved 45 minutes a
11:13day and helped avoid $12,000 a month in
11:15scheduling errors. The crew didn't have
11:17to change a single habit, and that's the
11:18power [music] of meeting a business
11:20exactly where they already work. They
11:21don't want to learn new processes, they
11:22don't want to learn a new UI, they just
11:24want to be faster. And I'm not even
11:25going to list industries for this one
11:27because honestly, every single business
11:28has this problem. If a business has more
11:30than a few employees and uses more than
11:31just one software tool, then they need
11:33reporting. And this sells because the
11:34definition of a flywheel that I always
11:36talk about. You design a system that
11:37basically saves the business [music]
11:38time. It's lets them serve more clients,
11:40serve existing clients better, and all
11:42of this grows the business. And when the
11:43business grows, the system gets used
11:44more, which basically feeds right back
11:46into itself. So, the value is
11:47exponential and that's real leverage.
11:49Okay, so now that you know what the five
11:51best workflows are to sell, here's how
11:52to actually sell them. Because I see so
11:53many people start out trying to sell the
11:55actual workflow. You need to be 10 hours
11:57a week, cut your
12:00admin mistakes and human error, speed up
12:02your leads. That positioning is what
12:03people actually pay for. So, instead of
12:05offering to build an end-to-end workflow
12:06or a cloud code automation, you're
12:08actually offering to save time, cut
12:09costs, help them focus on making more
12:11money. And there's actually a simple way
12:12to start doing this. Don't try to learn
12:13all five of those automations. Just pick
12:15one, learn it inside and out, and build
12:17a simple version of it. Then, show a
12:18business owner a demo. Now, once you
12:20pick your lane, you've got two paths
12:22that you can take. So, path one is to
12:23completely niche down on a specific
12:25process. You basically say, "Hey, I'm
12:27going to be the absolute expert on
12:28speed-to-lead systems. That's my [music]
12:30thing." And there's a huge benefit to
12:31this, which is you learn the language of
12:33that specific problem. You learn what
12:34converts, you learn what objections come
12:36up, you learn where things typically
12:37fall through the cracks in that process,
12:39and you legitimately become the expert,
12:41which means you have better case
12:42studies, you have better positioning,
12:43you have stronger testimonials, and you
12:45can likely promise better results than
12:47someone over here who's kind of more of
12:48a generalist and has 10 different
12:50services. Same reason why a really good
12:52steakhouse doesn't need hot dogs because
12:54they are known for their steak. And it
12:55also means that they can charge way more
12:56for their product. So, then the second
12:58path is to kind of go more broad and be
13:00a consultant. And if you go this route,
13:02your job isn't to walk in with your demo
13:04solution and force it on a business,
13:05your job is to understand all of these
13:07use cases so that you can make smart
13:09suggestions. But ultimately, you're
13:10there to identify the actual bottleneck.
13:12This isn't what the business owner
13:13thinks the problem is or what he or she
13:15thinks the AI solution that he or she
13:17wants is, it's actually finding the
13:18constraint. And there's a really good
13:19question that you can use to find it,
13:20which is saying, "Hey, if 500 new
13:22clients showed up at your business
13:24tomorrow, what would break first?" And
13:25this is super powerful because that
13:26question forces them to think through
13:28their entire operation logically, step
13:30by step. And as they walk through it,
13:31you'll find every single hole. Maybe
13:32their intake process falls apart. Maybe
13:34their team can't follow up fast enough.
13:35Maybe their reporting breaks down and
13:37nobody knows what's happening. Whatever
13:38breaks first, that's probably where you
13:39want to start. Think of it like this.
13:41Imagine your business as a pipe and you
13:42have water flowing through it and that
13:44is basically your cash flow. That's your
13:45money. If there's a clog near the
13:47beginning of the pipe, then why in the
13:49world would you try to pour more water
13:51into this clogged pipe? But that's what
13:52most businesses do. They spend more on
13:54ads, they spend more on sales people,
13:55and they do all of this stuff to try to
13:57put more volume through a system that's
13:58already broken. What you should actually
14:00be doing is removing that clog. And
14:01then, once it's gone and the water can
14:03go through the pipe, then you focus on
14:05finding more water and putting more
14:06water through it. And the five workflows
14:08that we just covered are the most common
14:10clogs. Speed-to-lead fixes the response
14:11gap. Document processing fixes the
14:13operational bottleneck. Follow-up
14:14sequences fix the leaky pipeline.
14:16Database reactivation unlocks revenue
14:18that's already been paid for. And
14:19internal reporting [music] fixes
14:20visibility gap that slows down every
14:22decision. So, pick your path, pick your
14:24first workflow, and go show someone the
14:25value. That's how this starts. So, I
14:27know that we covered a ton of
14:28information in this video. So, what I
14:29did is I threw all of this into a
14:31resource guide that you can access for
14:32completely free in my free school
14:34community. The link for that will be
14:35down in the description. And if you want
14:36to see exactly how I started making
14:37money using this framework, then check
14:39out this video on screen right up here.
14:40But anyways, that's going to do it for
14:41today. If you guys enjoyed or you
14:43learned something new, then please give
14:44a like. It definitely helps me out a
14:45ton. And as always, I appreciate you
14:46guys made it to the end of the video.
14:47I'll see you on the next one.
14:49Thanks, everyone.

About this transcript

This page contains the full transcript of I’ve Built 500 AI Workflows, This is What Businesses Want in 2026 by Nate Herk | AI Automation, generated from the public captions YouTube serves with the video. The transcript has 3,812 words across 575 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.