YouTube2Text

Every AI Business Ranked from Worst to Best — Transcript

by Nate Herk | AI Automation · 3,574 words · 507 segments · language en · Watch on YouTube

Full transcript

  1. 0:00So, there are a ton of businesses that
  2. 0:01you can start with AI and every day more
  3. 0:03and more are being created. That's why
  4. 0:05choosing the right business to start is
  5. 0:07harder than ever. Now, I've run some of
  6. 0:08these businesses, I know friends who run
  7. 0:10them, and I see a lot of my community
  8. 0:11members start multiple businesses on
  9. 0:13this list. So, I've seen and know these
  10. 0:15businesses [music] inside out. So, in
  11. 0:16this video, I'm going to rank each AI
  12. 0:18business that you can start from worst
  13. 0:20to best. I'm going to rank them based on
  14. 0:22how technical you need to be, the market
  15. 0:23space, and long-term durability. So, by
  16. 0:26the end of this video, you'll know which
  17. 0:27AI business you should try to build and
  18. 0:28which ones to ignore. So, a quick
  19. 0:30preface, every business on this list can
  20. 0:33work. It's really just a matter of what
  21. 0:34types of problems that you want to have.
  22. 0:36It's obviously based on my experience
  23. 0:38and what I've seen, so it's a little bit
  24. 0:39subjective. But, each one is going to
  25. 0:41get three scores out of five. Technical
  26. 0:43is how much skill that you need to
  27. 0:44deliver something that people will pay
  28. 0:46for, and a lower score on this ranking
  29. 0:48means more skill required. Market is
  30. 0:50whether there are more buyers than
  31. 0:51sellers right now, and durability is
  32. 0:53whether it can turn into a real business
  33. 0:55that lasts and keeps growing. Also, any
  34. 0:57prices that I mention are what
  35. 0:58businesses charge, not a promise of what
  36. 1:00you're going to make. But, with all that
  37. 1:02out of the way, let's start at the
  38. 1:03bottom of the list and work our way up.
  39. 1:04All right, so number 13 on the list
  40. 1:06might be a hot take, but AI SaaS, where
  41. 1:08you build software with something like
  42. 1:09Cloud Code, and you charge people a
  43. 1:11monthly subscription to use it. Now,
  44. 1:12this is the most hyped up one on the
  45. 1:14list, and to be fair, it also is one of
  46. 1:16the most scalable, but it's the hardest
  47. 1:17to start. Most people who successfully
  48. 1:19scale SaaS already had an audience or a
  49. 1:21service business or money going in in
  50. 1:23some way, because building the app might
  51. 1:25take an afternoon or a few days, but
  52. 1:27getting people to find it, pay for it,
  53. 1:28and to not churn out or to not cancel
  54. 1:30their subscription is the actual
  55. 1:31business, and it's not easy. So,
  56. 1:33technical here is a four, because
  57. 1:34security and infrastructure are going to
  58. 1:36get real once you have people actually
  59. 1:38using it, hundreds of people, thousands
  60. 1:39of people. Market is a one, roughly
  61. 1:41560,000 new apps hit the App Store in
  62. 1:44the first half of this year, which is
  63. 1:45almost as many as all of 2025, and
  64. 1:48downloads barely actually moved. And
  65. 1:50then, durability here is a three,
  66. 1:52because a moat can last, but one
  67. 1:53platform update could kill a thin
  68. 1:55wrapper. The ceiling is pretty huge,
  69. 1:57scalable with a small team, but it's the
  70. 1:58worst place to start in my opinion. And
  71. 2:00moving on to number 12, we have faceless
  72. 2:02AI channels. That could be YouTube,
  73. 2:03TikTok, or Instagram accounts where AI
  74. 2:06is making the content and you never show
  75. 2:07your face. Now, these can work, but
  76. 2:09views alone are not going to pay you.
  77. 2:10So, you have to be very thoughtful about
  78. 2:12what you're building. YouTube won't
  79. 2:13monetize mass-produced repetitive AI
  80. 2:15slop. So, you need an audience that
  81. 2:16somebody actually wants to pay to reach
  82. 2:19in the form of sponsorships. So, take a
  83. 2:20news page like PewDiePie, for example.
  84. 2:22They could theoretically automate that
  85. 2:23entire account with AI. And now they
  86. 2:25have reach, so sponsors will pay to get
  87. 2:27in front of it to get a post. A sports
  88. 2:28picks account, they could also sell
  89. 2:30people into like a paid Discord where
  90. 2:31people in there are also giving out more
  91. 2:33sports picks. So, brain rot content
  92. 2:35could go viral, but nobody's really
  93. 2:36paying to reach that audience. Now, I'm
  94. 2:38actually myself, kind of for fun,
  95. 2:40building a few faceless channels for AI
  96. 2:41Automation Society right now, and I'm
  97. 2:43thinking about the audience. And picking
  98. 2:45that comes before everything else. But
  99. 2:46this one really doesn't take much to get
  100. 2:48started as far as like capital or money,
  101. 2:50besides maybe just like a Cloud Code
  102. 2:52subscription. So, it can be fun to
  103. 2:53experiment with and maybe over time it
  104. 2:55grows into a real is going to be a two.
  105. 2:56Durability is going to be a one because
  106. 2:58the platform can change the rules on you
  107. 3:00overnight. But there is so much value in
  108. 3:02actually being able to have a channel or
  109. 3:03an account
  110. 3:07where you have a lot of people who
  111. 3:08actually care about that content. And if
  112. 3:10sponsors care to get in front of it,
  113. 3:11very, very valuable. All right, coming
  114. 3:13in number 11, AI websites. You basically
  115. 3:15just build sites for businesses with AI
  116. 3:17like Cloud Code or Codex or Lovable. So,
  117. 3:20it really just takes an afternoon
  118. 3:21instead of weeks. Now, this is one that
  119. 3:23I'd tell a total beginner to start with.
  120. 3:25I think it's a great place to start.
  121. 3:26It's easy, it's super visual. You can
  122. 3:28build a demo site in like 20 minutes
  123. 3:29before you even pitch to a business. A
  124. 3:31lot of dev agencies are still charging
  125. 3:335K to 10K for something that you can
  126. 3:35build in an afternoon because of AI,
  127. 3:37because those dev agencies might be
  128. 3:38doing it the old way where they're
  129. 3:39writing custom code. So, right now you
  130. 3:41can undercut them. But there is a
  131. 3:42ceiling to this business model. I think
  132. 3:44it's going to be hard to grow past, you
  133. 3:45know, 10, 20K a month, and AI site
  134. 3:48builders already sell for maybe a nine
  135. 3:50bucks a month subscription. So,
  136. 3:52technical here is a one, market here is
  137. 3:54a two, and durability is a one. I would
  138. 3:56say you use this when you're getting
  139. 3:57started, use it for momentum, and then
  140. 3:59move up the list. Number 10 though is AI
  141. 4:01operating systems, and that's one
  142. 4:03central setup where a company's agents,
  143. 4:05automations, tools, team knowledge all
  144. 4:08live inside of this one thing. Now, I
  145. 4:09love these as a foot in the door offer
  146. 4:11because once you earn their trust, you
  147. 4:12can easily win bigger projects like team
  148. 4:15brain setups or consulting projects,
  149. 4:17things like that. The problem with
  150. 4:19AIOS's is that it's hard to tie a KPI to
  151. 4:22it. And I think that every offer that
  152. 4:23you have needs one clear metric that
  153. 4:25you're moving, and this needs to be a
  154. 4:27business outcome metric, something that
  155. 4:28they care about. It also only works if
  156. 4:30the business owner, or whoever you're
  157. 4:31talking to, is already bought into the
  158. 4:33idea of AI. Otherwise, they won't really
  159. 4:35understand what they're paying for with
  160. 4:36this AIOS, and it might be hard to get
  161. 4:38them to actually habit shift to using
  162. 4:40it, and ultimately that means you might
  163. 4:41not get a great case study or any
  164. 4:43referrals out of it. So, technical here
  165. 4:44is a three, market is a four, and
  166. 4:46durability is going to be a two. And
  167. 4:48number nine is GEO, or generative engine
  168. 4:51optimization. It's basically SEO, but
  169. 4:54for AI. So, you're getting a brand
  170. 4:55recommendation when someone asks ChatGPT
  171. 4:57or Perplexity or Claude Code, whatever,
  172. 4:59a question. And money is pouring into
  173. 5:01this. Money's already been pouring into
  174. 5:02SEO. This is a proven model. Profound,
  175. 5:05one of the main tools for this, just
  176. 5:07raised $180 million at a $1.8 billion
  177. 5:10valuation. And only a handful of people
  178. 5:12in my community are even talking about
  179. 5:13this whole GEO business model. If you
  180. 5:15already do SEO, this is a great fit
  181. 5:17because you know how to prove results.
  182. 5:20And Vercel CEO says that ChatGPT is
  183. 5:22sending them about 10% of their new
  184. 5:23sign-ups, which is pretty big. Without
  185. 5:25that background though in SEO, [music]
  186. 5:27it can be hard to get results. And now
  187. 5:29that ChatGPT is also running ads, OpenAI
  188. 5:31may have more control in the future on
  189. 5:33things like visibility. So, technical is
  190. 5:34going to be a two here, market is a
  191. 5:36three, durability is a two. Okay, so
  192. 5:38number eight is AI receptionists and or
  193. 5:40support agents. That's AI that answers
  194. 5:42the phone, the website chat, DMs, emails
  195. 5:45for a business. [music]
  196. 5:46It can help book appointments. It can
  197. 5:48hand off to a human when it needs to,
  198. 5:49things like that. Now, the ROI here is
  199. 5:51super easy to explain. One study of 70
  200. 5:53million business calls found that only
  201. 5:55[music] 56% get answered by a person.
  202. 5:57And for a dentist or an HFA company, a
  203. 5:59missed call is a lost job. That's missed
  204. 6:01money. Set up for these can usually run
  205. 6:03like 1.5k to 5k plus a monthly retainer.
  206. 6:06But the catch here is durability.
  207. 6:08Podium, ServiceTitan, Weave, they're all
  208. 6:10building AI receptionists right now into
  209. 6:12their software that these businesses are
  210. 6:14already maybe paying for. So, technical
  211. 6:16here is going to be a three, market is a
  212. 6:17four, and durability is a two. Now,
  213. 6:19number seven is AI automation and
  214. 6:21integration. So, basically connecting AI
  215. 6:23into the systems that a business already
  216. 6:24runs, like their CRM, email inbox,
  217. 6:27invoicing. You can use tools like
  218. 6:28end-to-end, make.com, Cloud Code,
  219. 6:30whatever you want to use. This is where
  220. 6:32a lot of us started, me included. And
  221. 6:34demand for this kind of stuff is still
  222. 6:35growing, growing very fast, and it will
  223. 6:37not stop growing for a while. AI
  224. 6:38integration work on Upwork went up 187%
  225. 6:42in the last year. But the issue here is
  226. 6:43that just building automations is
  227. 6:45getting so much easier, so it's getting
  228. 6:46extremely commoditized. And businesses
  229. 6:48are now hiring for this kind of stuff
  230. 6:49internally. The skill that's actually
  231. 6:50worth paying for here is figuring out
  232. 6:52the bottleneck, tying the fix to a
  233. 6:53business result, and charging for that.
  234. 6:55And that's a different business model
  235. 6:56that I'm going to talk about later in
  236. 6:57this list, because now a single workflow
  237. 6:59can just go for like 500 to 2k. So,
  238. 7:01technical here is going to be a two,
  239. 7:03market is a three, durability is a two.
  240. 7:05By the way, guys, I've got this
  241. 7:06completely free SOP for you about
  242. 7:08getting your first AI automation client.
  243. 7:09It's going to go over the exact steps
  244. 7:11that has been proven for hundreds of our
  245. 7:12AI S+ members to get their first paid
  246. 7:15gigs. It goes over the one-sentence
  247. 7:16service pitch that can get you started
  248. 7:17today, why your first client should cost
  249. 7:19you money, the five-minute video that
  250. 7:21answers can this person actually deliver
  251. 7:22before you've actually received any
  252. 7:24money, what to do when you have zero
  253. 7:25case studies. There's so many good
  254. 7:27things in here that are going to help
  255. 7:29you out, even if you already do have
  256. 7:30clients, I would recommend grabbing this
  257. 7:32because like I said, it's yours
  258. 7:33completely free. So, if you want to grab
  259. 7:35this, there's a link for it down in the
  260. 7:35description. Let's get back to the
  261. 7:37video. Okay, so number six, AI
  262. 7:38workshops. [music] So, getting paid to
  263. 7:40teach a company's team how to use AI in
  264. 7:42their actual jobs. Published rates that
  265. 7:44I found go from $1,000 for a half day up
  266. 7:46to 5K for a full day workshop with a
  267. 7:48bigger team, and only 36% of employees
  268. 7:51say that they feel properly trained on
  269. 7:53AI, and I would bet that the real
  270. 7:54number, if you think outside of just
  271. 7:56that survey, is much lower than 36%.
  272. 7:58Now, I've done a bunch of these AI
  273. 7:59workshops myself, and it's a great great
  274. 8:02foot in the door. It builds trust
  275. 8:03instantly because you're in the room
  276. 8:04teaching. You see all the problems they
  277. 8:06have, and almost always after you do a
  278. 8:08workshop, do you win a bigger project
  279. 8:10because now they think of you as the AI
  280. 8:12expert in their lives. So, for this
  281. 8:13offer, technical is a three because
  282. 8:15teaching is a skill, and of course, you
  283. 8:16need to learn the tools that you're
  284. 8:18teaching. Market is going to be a three,
  285. 8:19and durability is a three. Okay, number
  286. 8:21five is interesting, and this is AI lead
  287. 8:23gen. You're generating leads for a
  288. 8:24business. You use AI to scrape and
  289. 8:26enrich and personalize and test way more
  290. 8:28outreach than a human team ever could.
  291. 8:30And I love this one because it's so
  292. 8:32sticky. If you're sending someone leads,
  293. 8:34and they're obviously monetizing those
  294. 8:35leads, they're not going to want to
  295. 8:36cancel the retainer if you're good at
  296. 8:38it, at least. And it's the clearest KPI
  297. 8:39on this whole list. More leads typically
  298. 8:41equals more customers or more revenue.
  299. 8:43Clay, one of the biggest tools in this
  300. 8:45sort of like AI lead space, passed $100
  301. 8:47million in annual revenue last year. But
  302. 8:49this business model isn't easy. You have
  303. 8:51to crack lead gen for your own business
  304. 8:53first because if you can't even generate
  305. 8:54your own leads, how would you be able to
  306. 8:56generate leads for someone else? And how
  307. 8:57would you get someone to trust that you
  308. 8:58can generate leads for them? The best
  309. 9:00spot to be in is an agency that already
  310. 9:02does lead gen, and then you just kind of
  311. 9:03add AI on top of it, and now you're an
  312. 9:05AI lead gen agency. So, technical here
  313. 9:08is a four, market is a three, but
  314. 9:10durability is a five. I think it's super
  315. 9:12super sticky, and everyone's always
  316. 9:13going to need leads. That's how
  317. 9:14businesses work. Now, number four, also
  318. 9:16very interesting, AI creative studio.
  319. 9:18Ads, commercials, carousels, video
  320. 9:20editing for brands. AI video editing was
  321. 9:22the fastest growing skill on all of
  322. 9:24Upwork last year, up 329%, and one
  323. 9:27person can now do what used to take a
  324. 9:28whole team and expensive gear. But
  325. 9:30anyone can just hit the generate button,
  326. 9:32even inside of Meta's own ad tools, and
  327. 9:34the average price for a UGC video
  328. 9:36dropped about 44% in just a year. So,
  329. 9:38what brands pay for is taste, knowing
  330. 9:40which model fits which audience, and
  331. 9:42running the ads and using the data to
  332. 9:44make the next ones. So, technical here
  333. 9:46is a three, market is a four, and
  334. 9:48durability is a three. This one's also
  335. 9:49very sticky if your creatives and if
  336. 9:51your ads and stuff like that actually
  337. 9:53work. So, it could be very very
  338. 9:55lucrative. Okay, number three is AI
  339. 9:57education and/or building a personal
  340. 9:59brand. So, [music] courses, communities,
  341. 10:01newsletters, or videos with your face on
  342. 10:03it or not, it's always better if your
  343. 10:04face is tied to it. And full disclosure,
  344. 10:06this is like the main business I run.
  345. 10:07That's kind of, you know, you guys are
  346. 10:08watching this video right now. So, take
  347. 10:10my score with that in mind. You have
  348. 10:11[music] to be in the weeds to teach this
  349. 10:13stuff, or at least be good on camera.
  350. 10:15And obviously this space is getting more
  351. 10:16competitive every day because there's
  352. 10:17more content creators popping up.
  353. 10:19There's also more people entering this
  354. 10:20market as consumers. So, it's a great
  355. 10:22opportunity. But, I think it is the most
  356. 10:23durable thing that you can build because
  357. 10:25when people are invested in you, you can
  358. 10:26pivot whenever the tools change. And the
  359. 10:28catch here though is speed. Building an
  360. 10:30audience takes years. And I did agency
  361. 10:32work and consulting and workshops long
  362. 10:34before I had a personal brand. Proof
  363. 10:36comes first and trust comes next. So,
  364. 10:38don't get distracted early. But, once
  365. 10:40you've found your lane, you can start
  366. 10:41spin-off businesses and you can power
  367. 10:42them with your brand, like a SaaS
  368. 10:44product later, or like, you know, all
  369. 10:46these other tools that I've mentioned
  370. 10:47earlier in this list, which is how all
  371. 10:48of the other businesses that I'm running
  372. 10:50right now are powered because of my
  373. 10:51personal brand and my education. So,
  374. 10:53this one is going to be a three for
  375. 10:55technical, a three for market, and a
  376. 10:57five for durability. Okay, moving on to
  377. 10:59number two now, AI consulting and
  378. 11:01audits. You go into a business, you
  379. 11:02figure out how it actually works, you
  380. 11:04find the bottleneck, and you prescribe
  381. 11:05where AI saves the most time or money.
  382. 11:07And so, this is like the consulting
  383. 11:09piece of the AI automation and
  384. 11:10integration that I talked about earlier.
  385. 11:12Think of it like being a doctor.
  386. 11:14Building the fix keeps getting cheaper.
  387. 11:15So, what they're really paying for is
  388. 11:17the diagnosis, plus all the human stuff
  389. 11:19like change management, super important.
  390. 11:21Governance, super important. It's
  391. 11:23getting stakeholders on board, culture,
  392. 11:25super important. LinkedIn ranked AI
  393. 11:26consultants the second fastest growing
  394. 11:28job in the US. And this year both
  395. 11:30Anthropic and OpenAI launched their own
  396. 11:32services companies to put AI into
  397. 11:34business. So now they have their own AI
  398. 11:36consultants that are implementing Claude
  399. 11:37or ChatGPT, things like that. Big
  400. 11:39investors are even buying up accounting
  401. 11:41firms and IT shops just to run them on
  402. 11:43AI. Now audits can usually be sold for
  403. 11:451K to 15K. So in this sort of market,
  404. 11:47technical is a three, market's a four,
  405. 11:49and durability's a four. Because AI can
  406. 11:51write an audit, but it can't run the
  407. 11:53change inside of a company and get
  408. 11:55people on board. And it beats education
  409. 11:57because it pays faster. And number one
  410. 11:59on this list is becoming a chief AI
  411. 12:01officer. Now there are two ways to do
  412. 12:03this, and I think they're equally good.
  413. 12:04You can become the AI person at one
  414. 12:06company that you already work at and
  415. 12:08become the chief AI officer and drive
  416. 12:09strategy and drive the P&L across the
  417. 12:11entire organization, or you can go
  418. 12:13fractional and you can be the AI lead
  419. 12:15for a few companies at once. And this
  420. 12:17can actually stem off of the one that we
  421. 12:18were just talking about, being an AI
  422. 12:20consultant, and then maybe you get
  423. 12:21promoted to like a fractional AI
  424. 12:22officer. Now I know the first one, being
  425. 12:24an in-house chief AI officer C-suite
  426. 12:26executive, isn't technically a business,
  427. 12:28but it can pay more than most businesses
  428. 12:30on this list. IBM surveyed 2,000 CEOs of
  429. 12:33big companies, [music] and 76% of their
  430. 12:35organizations now have a chief AI
  431. 12:36officer. And that was just out of 2,000
  432. 12:38large companies. But a year earlier,
  433. 12:40that number was 26%. PwC also found that
  434. 12:42workers with AI skills earn a 62% wage
  435. 12:45premium, and one recruiting firm puts
  436. 12:47full-time chief AI officer base pay at
  437. 12:49around 280K to 650K a year. Technical's
  438. 12:52a four because you need to know enough
  439. 12:54to never get caught out. Market's a
  440. 12:56five, durability's a five. And the more
  441. 12:57AI advances, the more valuable this role
  442. 13:00will become, I think. And I think it's
  443. 13:02going to become way more normalized for
  444. 13:03every organization to have a chief AI
  445. 13:05officer of some sort. So if you look at
  446. 13:07this whole list, the businesses at the
  447. 13:08bottom are where you're selling the
  448. 13:10thing you build, and the ones at the top
  449. 13:12are where you're selling your judgment.
  450. 13:13So if you're brand new, then start with
  451. 13:15something like websites to start getting
  452. 13:17some momentum. And then you want to use
  453. 13:18those same clients to work your way up
  454. 13:20towards things like workshops [music]
  455. 13:21and consulting and maybe even eventually
  456. 13:23something like that chief AI officer
  457. 13:24seat. But like I said at the beginning
  458. 13:26of this video, all of these can work.
  459. 13:28What matters is that you pick one and
  460. 13:29you focus on it like crazy. Instead of
  461. 13:31jumping to a new niche every few weeks
  462. 13:33or jumping business models every few
  463. 13:34weeks, just pick one and stay focused.
  464. 13:36Out of all of the entrepreneurs that
  465. 13:38I've talked to that have built
  466. 13:38successful businesses, it wasn't because
  467. 13:40they picked the best business at the
  468. 13:42right time and the best blah blah blah.
  469. 13:43It's because they had way more focus
  470. 13:45than everyone else who was trying to do
  471. 13:46that thing that they were doing and they
  472. 13:47stuck with it long enough to make it
  473. 13:49work. So pick one and make it work. So
  474. 13:51now you know which AI businesses you
  475. 13:53should build and which ones to avoid.
  476. 13:54But the window to build the best
  477. 13:55businesses won't stay open forever.
  478. 13:57Every day more and more people will
  479. 13:58learn about them and saturate the market
  480. 14:00and the ones who learn how to do this
  481. 14:01fast will be the ones that capitalize on
  482. 14:03this market. So if you want to start
  483. 14:04moving now, my free community has every
  484. 14:06course, every resource, every GitHub
  485. 14:08repo, skill, everything that I've shown
  486. 14:10on this channel for completely free. You
  487. 14:12can learn all the technical knowledge
  488. 14:13you need to build any of these
  489. 14:15businesses. Then if you want to take it
  490. 14:16further and you want to connect with
  491. 14:17other people who are actually building
  492. 14:18these different businesses and talk
  493. 14:19about, you know, clients and projects
  494. 14:21and scoping and things like that, then
  495. 14:23you can join my plus community. No
  496. 14:24matter where you're at, these
  497. 14:25communities will help you move fast. So
  498. 14:27you can join those communities with the
  499. 14:28links in the description, but that is
  500. 14:30going to do it for today. So if you guys
  501. 14:31enjoyed the video or you learned
  502. 14:32something new, please give me a like,
  503. 14:33helps me out a ton. And as always, I
  504. 14:34appreciate you guys making it to the end
  505. 14:35of the video. I'll see you all in the
  506. 14:36next one.
  507. 14:38Thanks, everyone.

About this transcript

This page contains the full transcript of Every AI Business Ranked from Worst to Best by Nate Herk | AI Automation, generated from the public captions YouTube serves with the video. The transcript has 3,574 words across 507 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.