YouTube2Text

The $1M+ Solo Grok Bot Business (Full Course) — Transcript

by Greg Isenberg · 9,018 words · 1,353 segments · language en · Watch on YouTube

Full transcript

  1. 0:00If you wanted to make money selling AI
  2. 0:02agents, how could you do it? What is the
  3. 0:04clearest path to building wealth selling
  4. 0:07AI agents? And what is the stack that
  5. 0:10you need? All the tools, all the skills,
  6. 0:11all the apps to put it all together.
  7. 0:14Well, today I brought on Nick from Orgo
  8. 0:16and he basically breaks down one of the
  9. 0:18opportunities I think is the most slept
  10. 0:20on. In fact, I don't I haven't seen
  11. 0:22anyone talk about how you can take a
  12. 0:24personal agent platform like Grokbot,
  13. 0:28niche it down, and sell it to people on
  14. 0:30a monthly basis. So today's episode is a
  15. 0:32full masterclass. It's under an hour and
  16. 0:35the people that stick around are going
  17. 0:36to understand all the techniques, the
  18. 0:38strategies, how to price it, how to
  19. 0:40think about this startup idea. It's
  20. 0:42literally a sauce factory. So uh I'm
  21. 0:45excited for you to get into it and by
  22. 0:47the end of the episode you're just going
  23. 0:49to really understand all the different
  24. 0:51parts, all the different tools. And I
  25. 0:52think that even if you don't want to go
  26. 0:54and build this startup idea, you're
  27. 0:55going to be way more AI native just by
  28. 0:58listening to this. Now if this episode
  29. 1:00gets more than 5,000 likes, comments,
  30. 1:02and subscribes, I'm going to include a
  31. 1:04link in the pinned comment [snorts] and
  32. 1:06that's going to have more ideas and more
  33. 1:07niches for this concept. So go, pause,
  34. 1:11go and do that, and then
  35. 1:13hit play again and I'll see you at the
  36. 1:15end of the episode.
  37. 1:17>> The startup ideas podcast.
  38. 1:20[music]
  39. 1:21It's sipping time, baby.
  40. 1:25>> Nick, by the end of this episode, what
  41. 1:27are people going to learn?
  42. 1:29>> [clears throat]
  43. 1:29>> People today are going to learn how to
  44. 1:31build their own custom Grokbot for a
  45. 1:34business and sell it for $25,000
  46. 1:37upfront, retain them on the back end
  47. 1:40with a down sell for like $3,000 a
  48. 1:42month. All in this video we will walk
  49. 1:45through how to build everything. How do
  50. 1:47you select the offer? How do you select
  51. 1:49the stack? How do you actually build the
  52. 1:50thing? We go through everything in this
  53. 1:52video and I can't wait for people to
  54. 1:55build this startup idea out. So stay
  55. 1:57tuned.
  56. 1:58>> So the commitment I need from you
  57. 2:00there's a full stack of different tools,
  58. 2:03agents,
  59. 2:04um open source technology. I need to
  60. 2:06understand everything about how I can go
  61. 2:09and start my Grok Grokbot for XYZ
  62. 2:13because I really do believe that um you
  63. 2:15know selling AI employees Grokbot for
  64. 2:18XYZ is one of the biggest opportunities
  65. 2:20right now. So I just need your
  66. 2:22commitment that if people stick around
  67. 2:24for this you know episode which will
  68. 2:26probably be under an hour um they will
  69. 2:29have the tactics, the techniques, the
  70. 2:32strategies to actually implement this
  71. 2:35uh today.
  72. 2:37>> Yes and you will see it. You see it
  73. 2:39right here. Like I have these agents
  74. 2:41here so you'll have examples and ideas
  75. 2:44of the types of agents you could build
  76. 2:46for businesses. You'll have examples and
  77. 2:48ideas of the types of apps you can build
  78. 2:50for them and you have a beautiful
  79. 2:52app already built for you that you could
  80. 2:54just fork and use for free and go sell
  81. 2:57it to a business ready for you all in
  82. 3:00this video. I walk through every step of
  83. 3:02the way.
  84. 3:03Um yeah there's nothing left out on this
  85. 3:05on this one.
  86. 3:06>> All right let's get into it.
  87. 3:07>> Let's do it.
  88. 3:09So I'm sharing my screen here. I want to
  89. 3:10show you this is a video of me. I'm
  90. 3:13going to mute it I made earlier this
  91. 3:15month
  92. 3:16and you still see this, right?
  93. 3:18>> Yep.
  94. 3:18>> I'm literally showing a video on Twitter
  95. 3:21of this tool called Bops that I built. I
  96. 3:23called it Bops cuz I'm like it's
  97. 3:25business ops and I wanted to you know
  98. 3:27build agents that solve business
  99. 3:29operation problems. And you you see here
  100. 3:31I'm like talking on the phone to the
  101. 3:32bots. Each bot has its own computer with
  102. 3:35four screens
  103. 3:36and I literally built this entire
  104. 3:39essentially Grokbot clone in a weekend.
  105. 3:43Um I posted on Twitter you know got 75k
  106. 3:46views and it just breaks my brain that I
  107. 3:50was able to do this in a weekend.
  108. 3:53And then so we have existing businesses
  109. 3:56and customers at at you know Orgo where
  110. 3:59we build them agents and I think that
  111. 4:01this is such a cool startup idea because
  112. 4:04it's an evolution of like
  113. 4:06um so many different ideas, which is
  114. 4:08like with AI
  115. 4:11there's a few ways that I personally
  116. 4:13have made money and and you can see my
  117. 4:15screen here, right?
  118. 4:16>> Yep.
  119. 4:17>> So I I've made money with AI by building
  120. 4:19apps. So the first when when Cursor came
  121. 4:22out, funny enough, fun story
  122. 4:24I watched the Greg Eisenberg podcast.
  123. 4:26You had Mickey on. He was teaching
  124. 4:28Cursor. I think this was like summer of
  125. 4:3024 I think, something like that. Um
  126. 4:34super early. Of course, you're always
  127. 4:35early to to everything and and I I built
  128. 4:38my first money-making business online
  129. 4:40building an app using AI. Then the
  130. 4:43second biggest way that I've made money
  131. 4:45from AI was with agents, selling agents
  132. 4:48to businesses, like selling a Hermes
  133. 4:50agent to a business for 5K a month. Now
  134. 4:52the interesting opportunity that we're
  135. 4:54at today is building agents and apps
  136. 4:59together
  137. 5:00and solving like the entire gap in a
  138. 5:02business with agents and apps. And so
  139. 5:06what why am I talking about apps? So
  140. 5:07Greg, have you heard about people who
  141. 5:09say
  142. 5:10apps are dead, software is dead?
  143. 5:12>> Most people on X are saying that.
  144. 5:15>> What what are your thoughts?
  145. 5:17>> Um my thoughts uh I actually just
  146. 5:20recorded like a solo episode about this.
  147. 5:23I think
  148. 5:24the people are saying that because
  149. 5:27uh they don't want other people to go
  150. 5:30and build things.
  151. 5:32So yes, do I think that software is
  152. 5:36evolving? 100%. Um
  153. 5:39but
  154. 5:40you know, I think that what's happening
  155. 5:42is the the era of like dumb tools, dumb
  156. 5:46SaaS where it's like you have a human
  157. 5:47going and like pressing buttons and
  158. 5:49uploading a CRM
  159. 5:50uh
  160. 5:51and uploading to a CRM. Like that is
  161. 5:53ending. I think we're moving more
  162. 5:55towards like
  163. 5:56the price for software is trending
  164. 5:58towards zero and you're going to give
  165. 6:00away the software and you're going to
  166. 6:02upsell agents and human intelligence to
  167. 6:06go as as like the premium experiences.
  168. 6:10That's my take on apps are dead, SaaS is
  169. 6:14dead. It's evolving and people don't
  170. 6:18want you to know that there's never been
  171. 6:19more opportunity in this space.
  172. 6:22>> Exact I wholeheartedly agree. And like
  173. 6:25in my perspective, it's like
  174. 6:28the era of generic apps
  175. 6:32is what to me feels like is dying.
  176. 6:35And the era of custom built like
  177. 6:38building your own CRM for your business
  178. 6:42that is like purpose built for you, like
  179. 6:45your taste, that is more alive than ever
  180. 6:48and it's only the beginning. And you
  181. 6:50know, that concept of like building that
  182. 6:53purpose built apps, purpose built agents
  183. 6:55for a business and for an industry, oh
  184. 6:58my gosh, like that's what I'm here
  185. 7:00talking about today when I'm like sell
  186. 7:02Grokbot to businesses, yes.
  187. 7:04Build it. Build your own Grokbot for an
  188. 7:07industry like insurance or marketing
  189. 7:10agencies or dental clinics. Like
  190. 7:13customize it so that the agents come
  191. 7:15pre-built into that out of the box. Each
  192. 7:18of them having their own phone number,
  193. 7:20email, um channel connector so you can
  194. 7:23talk to them on Slack, all of that which
  195. 7:24I'm going to show you how to build in
  196. 7:26this episode. Um and
  197. 7:28>> So what you're saying is basically, hey,
  198. 7:31you have all these big platforms like
  199. 7:32Grokbot, Dots, whatever, and they're not
  200. 7:36in the business of these niche
  201. 7:39setting up these like niche use cases
  202. 7:41for businesses. That's one piece and the
  203. 7:43other piece is the the the the owners of
  204. 7:46these businesses they don't know how to
  205. 7:49set it up and they don't even they're
  206. 7:51too busy anyways, right? They're just
  207. 7:53like operating their business. So you're
  208. 7:54saying, hey, there's an opportunity to
  209. 7:56actually spin up your own version
  210. 7:59custom. You've gone and validated that.
  211. 8:02You built it in a weekend.
  212. 8:04And and if you sort of add the right,
  213. 8:07you know, plugins, templates, it's going
  214. 8:09to be out of the box like a 10x
  215. 8:12experience for them because if they go
  216. 8:15and download Grokbot, if they go and
  217. 8:17download another one of these platforms,
  218. 8:19when when they unbox it, so to speak,
  219. 8:22they're going to be so overwhelmed,
  220. 8:23they're not going to know how to use it.
  221. 8:25And also like they might not want to
  222. 8:29like the idea of like calling like an AI
  223. 8:31transformation agency might even be too
  224. 8:33much for them.
  225. 8:34>> Exactly. Like they just want their
  226. 8:36problem solved. And and AI
  227. 8:38transformation agency, they're going to
  228. 8:39walk them through all of these
  229. 8:40processes. It's it's going to take them
  230. 8:42this much time. They're going to have to
  231. 8:43learn all these things. But if they
  232. 8:45could just install this thing into the
  233. 8:47business that is actively already
  234. 8:49solving all their problems, it has an
  235. 8:52outreach agent, it has a, you know,
  236. 8:54agent that, you know, manages their
  237. 8:55community on school, like all these
  238. 8:57different things. And I can kind of walk
  239. 8:59through what I use mine for. I have that
  240. 9:01here if you can see this.
  241. 9:03So like if you just had that that you
  242. 9:05could just install into the business,
  243. 9:07it's like they're willing to pay so much
  244. 9:08money for it. So I built Bops because
  245. 9:10actually I wanted to solve I didn't like
  246. 9:13how all the other
  247. 9:15agent platforms worked and I kind of
  248. 9:16wanted to marry some things I liked and
  249. 9:17some things I didn't like. And so I
  250. 9:19built it for myself and then I posted
  251. 9:20that video and some of our some some of
  252. 9:23our customers saw it and they're like,
  253. 9:25oh, like we want that. And so I can
  254. 9:27actually show you here like one of our
  255. 9:29insurance customers, like right now
  256. 9:31we're in the process of building them
  257. 9:32their own insurance agency Bops. And you
  258. 9:37can see here it's like custom branded.
  259. 9:38So like the computer here has their
  260. 9:40branding and it's like it's going to be
  261. 9:42specifically purpose built for
  262. 9:45insurance agencies.
  263. 9:47And so yeah, I can kind of show you like
  264. 9:51some of the cool things I'm doing with
  265. 9:52mine. It's it's really cool cuz what
  266. 9:54people don't realize is
  267. 9:56you can like all the same tools that
  268. 9:58these big companies are using to build
  269. 10:00Grokbot or Dots and whatever, they they
  270. 10:03make it available so that you can build
  271. 10:05your own version too. Like they they
  272. 10:06have their agents API and they have all
  273. 10:09the same, you know, the voice models and
  274. 10:11all of that available to build on. And
  275. 10:13so you can actually have a better
  276. 10:15product in a lot of ways than what
  277. 10:17they're offering cuz you can glue
  278. 10:19together just everything that you love
  279. 10:21across the internet of like, oh, I want
  280. 10:23an email from Agent Mail or I want a
  281. 10:25computer from Orgo or like this,
  282. 10:27whatever. And you can put it in your own
  283. 10:29combination of flavors and and really
  284. 10:31build your own better version of it. So
  285. 10:35um this is kind of cool. So I have my my
  286. 10:38Bappy is my chief of staff and I can
  287. 10:41email him. So if I click like his
  288. 10:43profile here, you can see he has an
  289. 10:45email. He has a phone number. I could
  290. 10:47text him. So if I text him like right
  291. 10:49now actually, let me say
  292. 10:51um
  293. 10:53I could say
  294. 10:54say hi to the podcast.
  295. 10:57And when I text him like it shows up
  296. 10:59here and says say hi to the podcast and
  297. 11:02he's going to
  298. 11:03reply and he's going to say hi. And then
  299. 11:05I could
  300. 11:05>> really cool.
  301. 11:06>> Yeah, and I could call him. So I'm going
  302. 11:08to call him on my phone. Like can Grok
  303. 11:11bot do this? I mean, I don't so.
  304. 11:12[clears throat] So I call him on my
  305. 11:13phone.
  306. 11:15>> Hello, can you hear me?
  307. 11:16>> Hey Bappy, yeah, I can hear you. Can you
  308. 11:18pull up um can you pull up like a list
  309. 11:22of companies
  310. 11:23uh that I could sell to and do pull up
  311. 11:27their websites as well and then also
  312. 11:30pull up the S&P 500. Let's take a look
  313. 11:32at that. And I also want you to
  314. 11:36keep a look at Hacker News to see any
  315. 11:38relevant posts uh for me today.
  316. 11:44And so I tell I tell him to do that.
  317. 11:46You you hear that okay, Bappy?
  318. 11:48And you can see over here, you you see
  319. 11:50on my screen, he's he's
  320. 11:52so it says what I say, so I'm going to
  321. 11:53hang up. So I I end the phone call and
  322. 11:55now he's going to go ahead and he's
  323. 11:56going to do all that. And so
  324. 11:59I literally that's a phone number. I
  325. 12:01called him on his phone number.
  326. 12:03And now look, he has his own computer
  327. 12:05and he has like four screens on this
  328. 12:07computer that he can control and each
  329. 12:09bot here has their own computer with
  330. 12:11four screens and he can multitask and he
  331. 12:13could do all these different things at
  332. 12:15once. And so like to me, this is like a
  333. 12:18superior product, and I was able to
  334. 12:20build it as one person. Now obviously I
  335. 12:22have my own experience and collection of
  336. 12:23ideas that I put together to get here,
  337. 12:25but like genuinely, I'll show you how to
  338. 12:28do this. Like Opus 5.5, you can build
  339. 12:31anything. Like you're 323 prompts away
  340. 12:35from building your anything.
  341. 12:38In my case it was it was bops in a in a
  342. 12:40weekend.
  343. 12:43So
  344. 12:44yeah, just a quick demo on that.
  345. 12:47>> Yeah, I think like
  346. 12:48it's so obvious that if you were to
  347. 12:51install a Grokbot for a business,
  348. 12:54especially like a SMB, they would want a
  349. 12:57phone number attached to the bop, the
  350. 13:00bops, not the the bops, the bots.
  351. 13:04You know, so that they're
  352. 13:06they're able just to quickly voice note,
  353. 13:08call them, text them. It's so obvious,
  354. 13:10right?
  355. 13:12And it just goes to show you how early
  356. 13:15we are in this whole personal agent era
  357. 13:17that it does feel in some ways like so
  358. 13:19hobbyist. You do have to like you you
  359. 13:22have to like find out how other people
  360. 13:24are using it and and then like copy it.
  361. 13:27You know what I mean? So if you I mean
  362. 13:29the thing here is you have to understand
  363. 13:31the ICP really well
  364. 13:33and and you have to understand their
  365. 13:35pain points, and you have to understand
  366. 13:37their bottlenecks, and you have to come
  367. 13:38with them with solutions.
  368. 13:41>> Exactly. Like you really need to exactly
  369. 13:45that. Like and you can see here I even
  370. 13:47have a managing my Twitter and
  371. 13:48everything.
  372. 13:49You need to understand the industry
  373. 13:50specific problems really well for you to
  374. 13:53even make agents that are useful. Like I
  375. 13:56have a bunch of agents here. I actually
  376. 13:59a quick tip for people. Like I
  377. 14:01personally build an agent for every
  378. 14:03platform. So I have a Twitter agent, I
  379. 14:06have an email agent, I have a school
  380. 14:10platform agent, and each agent owns this
  381. 14:13platform and is like logged in and is
  382. 14:16able to automate tasks and routines
  383. 14:18relevant to that platform for me. Uh so
  384. 14:21for instance, my LinkedIn agent can
  385. 14:23outreach to to uh prospects that are
  386. 14:27relevant for, you know, um for our
  387. 14:29industry and people that I might want to
  388. 14:31sell to and connect with them and then
  389. 14:34once it connects with them, then send a
  390. 14:36message after the connection
  391. 14:38and then set the booking uh appointment,
  392. 14:41get the calendar booked appointment,
  393. 14:43then I jump on the call. I have the
  394. 14:45brief already. Like that's how I use
  395. 14:47these things. I have a whole bunch of
  396. 14:48them here. Um and so
  397. 14:51yeah, that's just a quick tidbit on
  398. 14:53like, you know, how to kind of divvy up
  399. 14:55an agent for what use case. Um so
  400. 14:59that is kind of the uh hopefully the
  401. 15:02value proposition can be, you know,
  402. 15:04explained
  403. 15:05by just seeing that. And now I want to
  404. 15:08kind of walk through how the hell did I
  405. 15:10build this in a weekend? How can you do
  406. 15:12it too? What is the kind of offer that I
  407. 15:14would take to a business owner to charge
  408. 15:17$25,000 up front? You know, maybe you
  409. 15:19collect half of that up front and then
  410. 15:21the other half at the end of the
  411. 15:22implementation of this. Um but let's go
  412. 15:25through the entire uh build out process,
  413. 15:28shall we?
  414. 15:29>> Let's do it.
  415. 15:30>> So first things first, I want to kind of
  416. 15:33walk through
  417. 15:34how I build in general cuz it 's a
  418. 15:37principal thing. You don't need to think
  419. 15:40too much about prompt engineering these
  420. 15:42days. When you use something like Opus
  421. 15:455.5 on high mode, I mean it just the
  422. 15:48model is just so smart and intuitive
  423. 15:51that it kind of just gets it. So um but
  424. 15:54there are principles that allow me to
  425. 15:57build well and fast and I would like to
  426. 16:00walk through those. So the first things
  427. 16:02first
  428. 16:03is um well obviously you actually need
  429. 16:06first things first to decide well how if
  430. 16:08I'm going to build something like where
  431. 16:10am I going to build it in? So you can
  432. 16:11build inside of Claude code. You can
  433. 16:13build inside of uh T3 that's this Theo
  434. 16:16got on Twitter. He has this T three
  435. 16:17code.
  436. 16:18I actually built Bops inside of T three.
  437. 16:21And I'll show you here.
  438. 16:22Uh this is the the the exact
  439. 16:35of T three.
  440. 16:38I I'm not even kidding like I did this I
  441. 16:40built this like an idiot. Like I
  442. 16:41literally just kept going in one single
  443. 16:44long thread.
  444. 16:45I didn't even like connect it to GitHub
  445. 16:47or anything. I didn't push like I just
  446. 16:50did one long thread for 36 hours. I had
  447. 16:52loops running overnight.
  448. 16:54And that is how I built this entire
  449. 16:56thing. So you use something like T
  450. 16:58three. You open up a project. You select
  451. 17:01your model like Opus 5.5.
  452. 17:04And the the process is I always start
  453. 17:07with research. So if you look at the
  454. 17:10very first message I sent inside of
  455. 17:13building this this project of Bops,
  456. 17:16I literally was like
  457. 17:17I want to build a version of Grokbot or
  458. 17:20Meta's Muse or OpenAI's Dots very
  459. 17:22similar. Research them so you understand
  460. 17:25first.
  461. 17:26And so what it does is it goes ahead it
  462. 17:29researches all of them to conceptually
  463. 17:31understand it. And so what I'm trying to
  464. 17:32do here is I'm trying to
  465. 17:34frame and prime the model to understand
  466. 17:37what am I even talking about? What is it
  467. 17:39the thing that I'm even trying to build
  468. 17:41to begin with?
  469. 17:43And so it researches all of these. It
  470. 17:45has its sources here so I know that the
  471. 17:47model knows what I'm talking about. And
  472. 17:49then the next thing I do is I design. So
  473. 17:54have you you're are you familiar with
  474. 17:55Paper, Greg?
  475. 17:57>> Big fan. Not affiliated but big fan of
  476. 17:59it.
  477. 18:00>> I'm also not affiliated like the only
  478. 18:02thing I'm affiliated with all these
  479. 18:03tools I'll tell you is is Orgo.
  480. 18:05Obviously I'm a co-founder of Orgo so we
  481. 18:07we build the computers so that you could
  482. 18:08build your own Grokbot. So if you want
  483. 18:10to build your own Grokbot you could use
  484. 18:11our computers. But Paper
  485. 18:13is like Figma for the AI era.
  486. 18:17And I designed everything inside of
  487. 18:20Paper. You simply give uh your coding
  488. 18:24agent, so inside of T3, I have Opus 5.5
  489. 18:26and I installed the Paper MCP. And what
  490. 18:29that looks like is this.
  491. 18:31And so I I did that research step first.
  492. 18:34Now my agent knows what what I'm trying
  493. 18:36to build. And then the second thing I
  494. 18:38did was I said, okay, let's design this.
  495. 18:41Before building anything, I want to
  496. 18:43design inside of Paper what this would
  497. 18:47look like. And so it comes up with these
  498. 18:49designs.
  499. 18:50And I'm like, okay, okay, and I just
  500. 18:53kept going. I kept like back and forth,
  501. 18:55back and forth. I wanted to look like
  502. 18:56this, plain English, plain you know,
  503. 18:58natural language. I'm telling it what I
  504. 19:00want. And you can see the evolution of
  505. 19:04how this is all built out. This is
  506. 19:06before any building, just all design. Is
  507. 19:08this Is this adding up?
  508. 19:10>> It's crazy, dude. No, it's it's just
  509. 19:12it's insane because
  510. 19:15the also like the the branding is so
  511. 19:17good. And I feel like Paper Paper
  512. 19:19actually low key excels at at branding.
  513. 19:23>> I don't know what they did to do that,
  514. 19:25but it's genuinely true. Like
  515. 19:28they I don't know what they added in the
  516. 19:30MCP or their app or whatever. Like
  517. 19:33anything with Opus 5.5 and designing
  518. 19:36inside of Paper comes out like an S-tier
  519. 19:38product design. Like it's insane.
  520. 19:41Um and so yeah, you can see like this
  521. 19:43whole like it's a long process. I mean,
  522. 19:45I I spent a lot of time thinking of like
  523. 19:48the simple design. And you can see I
  524. 19:50landed on this like
  525. 19:52Apple-esque, you know, like you're
  526. 19:54familiar in iMessage and FaceTime and
  527. 19:56all of this. Like I landed on that kind
  528. 19:58of design, very you know, inspired by
  529. 20:01Grokbot as well. Um and that's before
  530. 20:04any building. So
  531. 20:06that's the second step. So first step,
  532. 20:09we research. Second step, we design.
  533. 20:12>> Quick question on research and design.
  534. 20:15>> Yeah.
  535. 20:16>> Um
  536. 20:17T3 code,
  537. 20:18why did you use it versus say Claude or
  538. 20:22or or something.
  539. 20:25And any tips there? And then my second
  540. 20:27question being Paper.
  541. 20:29Uh obviously we like Paper, but like
  542. 20:32what are some best practices in terms of
  543. 20:35using Paper
  544. 20:36to get the most out of it?
  545. 20:38>> Yeah. So I chose T 3 genuinely out of
  546. 20:43curiosity. I was like I had seen this
  547. 20:46guy on Theo. So yeah, if you if you guys
  548. 20:48aren't making content yet for your
  549. 20:49product like do it because eventually
  550. 20:51people will see you enough when you make
  551. 20:53so much content that they'll try your
  552. 20:54thing. That's how I tried T 3. Um and I
  553. 20:57actually so my take on it is it's very
  554. 21:01good. It's very simple and you can see
  555. 21:03it's a very simple design.
  556. 21:05So I I recommend it. It's free and if
  557. 21:08you already have a Claude subscription,
  558. 21:10you could just use it inside of T 3.
  559. 21:11It's free.
  560. 21:12And I think it's open sourced. So
  561. 21:15um yeah, I just
  562. 21:17I like things that work and are simple
  563. 21:19and and they don't get in my way, you
  564. 21:20know? Um so that's my take on T 3. And
  565. 21:24then as far as Paper, to get the most
  566. 21:25out of it
  567. 21:27this is a principle
  568. 21:29uh it's a design thinking principle
  569. 21:31which is
  570. 21:32um diverge and then converge. And so
  571. 21:35what I mean by that is inside of Paper
  572. 21:37you'll see I many times will ask it to
  573. 21:40generate me like multiple variations.
  574. 21:43You can see here multiple variations of
  575. 21:46this same kind of feature idea that I
  576. 21:48have in mind. And
  577. 21:51I ask it to give me you know three four
  578. 21:52variations and then that's the diverge
  579. 21:55part and then I converge and I say okay,
  580. 21:57I like this part from choice one and
  581. 21:59this part from choice three. And then
  582. 22:01that's how you're able to really
  583. 22:03get the most out of something like Paper
  584. 22:05or any just design process in general.
  585. 22:07Diverge and then converge. There's a
  586. 22:09a hack there.
  587. 22:11Um
  588. 22:12so going further along into the stack
  589. 22:15um after I research, after I design you
  590. 22:18have to choose now what tools you're
  591. 22:21going to use to build the thing. So for
  592. 22:23Bops there's a few different components
  593. 22:26and it's worth kind of like thinking in
  594. 22:28this like in this way of uh
  595. 22:31just like as an engineer, you know,
  596. 22:34engineer thinking of okay,
  597. 22:36I have a phone, I have a computer, I
  598. 22:40have an email. Like where are all these
  599. 22:41different things coming from? And so you
  600. 22:43have to decide. For me, I wanted to keep
  601. 22:46it simple. I didn't want to build this
  602. 22:48infrastructure myself. So obviously like
  603. 22:51I use Orga. I'm a co-founder of Orga. We
  604. 22:53build the computers so you don't have to
  605. 22:54build them yourself. You can use them to
  606. 22:56build your own bops. Uh I used Agent
  607. 22:59Mail to give it an email. I used Agent
  608. 23:02Phone to give it a phone number. I used
  609. 23:05Composio to give it these connectors. So
  610. 23:08you can see here when I go to uh logins
  611. 23:11and apps, I'm sorry, apps is the
  612. 23:14the bots can actually make, you know,
  613. 23:16apps. When I go to logins here, I can
  614. 23:19add a connector
  615. 23:21and I have all these connectors out of
  616. 23:23the box from this tool called Composio.
  617. 23:27Literally just one one API connection to
  618. 23:30Composio. I get all of these connectors
  619. 23:32for my application. And then I can
  620. 23:34connect to them from here. So all my
  621. 23:36bots can have access to that.
  622. 23:38Um so
  623. 23:40deciding the stack
  624. 23:42that you're going to use after you
  625. 23:43design it is is the next part. So okay,
  626. 23:46I always tell it to read the docs and
  627. 23:49read the API references. Those are the
  628. 23:50two things I do with every kind of
  629. 23:52feature. So I'm like okay, if I'm going
  630. 23:54to build the Orga computers instead of
  631. 23:55bops,
  632. 23:56read the API references for Orga and
  633. 23:59read the docs. That's what I tell my
  634. 24:01agent to do.
  635. 24:03Next up,
  636. 24:04after it builds it out, I literally
  637. 24:08Do you use CleanShot? What do you what
  638. 24:10what's your what's your screen You use
  639. 24:11CleanShot? Okay.
  640. 24:12>> I've been on it for years. Also not
  641. 24:14affiliated, but love it.
  642. 24:16>> You strike me as a CleanShot guy, Greg.
  643. 24:18>> I mean it it really it's one of those uh
  644. 24:22upgrades
  645. 24:24on your Mac that once you use it, you
  646. 24:26cannot go back.
  647. 24:28>> It's
  648. 24:28>> it's just simple. It just works, right?
  649. 24:31>> Yep.
  650. 24:32>> So, CleanShot for those who don't know,
  651. 24:34it's a screenshot app. It's just better
  652. 24:36than the native screenshot inside of
  653. 24:39your apps and inside of your Mac.
  654. 24:41So, I literally will screenshot, you
  655. 24:44know, the application as the agent is
  656. 24:46building it. If there's something I
  657. 24:47don't like or if there's a bug, I will
  658. 24:48screenshot it and I will send it to Opus
  659. 24:515.5 and T3
  660. 24:53and I will say fix this or I don't like
  661. 24:55this or can you make this more like
  662. 24:57that.
  663. 24:58And so screenshots are awesome. And many
  664. 25:00times I'll screenshot things from like
  665. 25:02GrokBot that I like. Like, oh, I like
  666. 25:04this UI element of GrokBot. I'll
  667. 25:06screenshot that. I'll send it to my
  668. 25:07agent. It'll implement that inside of my
  669. 25:09app.
  670. 25:10Okay, let's wrap.
  671. 25:12>> The screenshot bit is like in general,
  672. 25:14if you're building anything, you want a
  673. 25:15swipe file.
  674. 25:17A swipe file is a place where you have
  675. 25:20references, screenshots on copy and
  676. 25:22design, and then you'll feed those,
  677. 25:25you know, to
  678. 25:28to the LLMs or to paper to help get the
  679. 25:30most
  680. 25:31out of it.
  681. 25:32>> Yeah. Yeah, and I I if yeah, if you
  682. 25:36don't have a swipe file for like things
  683. 25:38that you it's just a life hack. You
  684. 25:40should really have a swipe file for
  685. 25:41everything. Like I'm talking marketing.
  686. 25:44If there's like posts that you like,
  687. 25:45that you get inspiration from or apps
  688. 25:47that have cool UIs that you you really
  689. 25:49get inspired from, it's good to get in
  690. 25:51the process of creating like Greg said,
  691. 25:53like these swipe files of just
  692. 25:55inspiration, like your own Pinterest
  693. 25:57feeds, you know.
  694. 26:00Last let's let's wrap up quickly on
  695. 26:02this. I don't want to spend too much
  696. 26:03time. The bot life
  697. 26:06I talk about that, you know, give every
  698. 26:07agent its own email, phone number,
  699. 26:10computer,
  700. 26:12things to make this thing these things
  701. 26:14actually useful. The the cool thing when
  702. 26:16you build it yourself, like inside of
  703. 26:18GrokBot, you don't actually get a new
  704. 26:21computer for all of the bots. I don't
  705. 26:22know if people know this, but in
  706. 26:24GrokBot, it's actually each bot gets its
  707. 26:26own screen, but it doesn't get its own
  708. 26:29computer. And so when you build stuff
  709. 26:31yourself, it's like you can give each
  710. 26:32bot its own computer for privacy reasons
  711. 26:35and just multitasking reasons and all of
  712. 26:37that.
  713. 26:38Um so
  714. 26:39give it the things that'll make it
  715. 26:41useful, power it up, boost it up, uh
  716. 26:44decide where you're going to host it.
  717. 26:45You can actually host it on Orgo. Orgo
  718. 26:46is like the computer, but it can act as
  719. 26:48the server. Uh you can host it on Fly,
  720. 26:51you can host it on Vercel.
  721. 26:52And uh finally just ship fast. Once
  722. 26:55again, this this took 36 hours.
  723. 26:59I made a video on it. I'm literally
  724. 27:01eating cereal in the video. So like this
  725. 27:04almost did as good as our YC launch
  726. 27:06video and it took me 2 minutes to make.
  727. 27:09So you don't need
  728. 27:11VC funding, you don't need a lot of
  729. 27:13money.
  730. 27:14You need a weekend and enough
  731. 27:16inspiration to go build something really
  732. 27:18freaking cool and just show the world.
  733. 27:21And in this case with like if you build
  734. 27:23your own Grokbot for an industry, you
  735. 27:25you have two options. Let's Let's dive
  736. 27:26into the how we're going to make money
  737. 27:28from this and what's the offer and
  738. 27:29what's the stack.
  739. 27:31There's one bucket which is like, okay,
  740. 27:33we can either sell this one to many
  741. 27:36or
  742. 27:38we can sell this
  743. 27:39um like one to few, which is like
  744. 27:42selling to the SMBs, right?
  745. 27:45Now Greg, if you were to sell this to
  746. 27:46SMBs, like I I mentioned we could sell
  747. 27:49this for like 25k upfront
  748. 27:52and probably do like a 3k a month
  749. 27:55retainer on the back end to maintain it.
  750. 27:57Would that be an offer that makes sense
  751. 27:59to you? Would you agree with that?
  752. 28:02>> Um I mean
  753. 28:05I you know, I'll play devil's advocate
  754. 28:07as if I'm the customer. I mean, let's
  755. 28:09say I own a dental clinic. Like my
  756. 28:12question to you would be like
  757. 28:14how am I going to be able to make, you
  758. 28:17know, $36,000
  759. 28:19plus 25k, $51,000 of value from
  760. 28:24from this?
  761. 28:25>> Yeah, I I would I would actually just
  762. 28:28say I would like I would pull up I would
  763. 28:30pull up the thing that I built you and I
  764. 28:31would look at one of the bots and I
  765. 28:33would be like, Greg,
  766. 28:35how much how much do you pay somebody
  767. 28:38like a sales rep to like in your case if
  768. 28:41you're a dental clinic, uh how much do
  769. 28:43you pay your your front desk lady to do
  770. 28:46all the email and support, all of the
  771. 28:48billing, all of the calendar bookings
  772. 28:50and everything and appointments and all
  773. 28:52of that. How much do you pay her a year?
  774. 28:54And you're probably going to say, well,
  775. 28:55I pay, you know, 50 or 60 or $70,000 a
  776. 28:58year for just that one person. And I'm
  777. 29:00going to show you. I'm going to look I'm
  778. 29:01going to say, look,
  779. 29:02there's that one person right here and
  780. 29:04that's just one of the 15 people, 15
  781. 29:08bots that I'm giving you here today
  782. 29:10upfront done
  783. 29:12ready to work.
  784. 29:13Uh and you own it and you own this
  785. 29:15system.
  786. 29:17And I think like that's the key is
  787. 29:19when you try um in the past like
  788. 29:24you can sell agents, right, for 5k a
  789. 29:26month as like a Hermes agent to a
  790. 29:28business. But there's a lot of frictions
  791. 29:31um that a lot of people run into with
  792. 29:33that model. And what I've seen work
  793. 29:35really well for a lot of people is like,
  794. 29:38oh, they don't business owners don't
  795. 29:39want to be feel like they're held by the
  796. 29:41neck on a retainer for these AI
  797. 29:43employees. So when you sell them this
  798. 29:45system and you're like, I'm going to
  799. 29:46just build this system, I'm going to
  800. 29:48release it into your business, it's 25k,
  801. 29:50you own it.
  802. 29:52And then you down sell them after that
  803. 29:53after a 30-day implementation and you're
  804. 29:55like, okay, cuz they're going to want
  805. 29:58your help to maintain and things like
  806. 29:59that.
  807. 30:01And that's completely optional. That's
  808. 30:03where the 3k a month comes in. But when
  809. 30:04you sell them this concept of like they
  810. 30:06own it and they can have it forever,
  811. 30:09that seems worth it to a lot of
  812. 30:11businesses.
  813. 30:12>> My take is here's how you can turn that
  814. 30:14offer to make it completely
  815. 30:16irresistible.
  816. 30:18So
  817. 30:19a lot of these companies that you know,
  818. 30:22we're talking about targeting, their
  819. 30:24websites
  820. 30:25aren't super optimized.
  821. 30:28And we've shown in this episode how you
  822. 30:30can create
  823. 30:32just gorgeous gorgeous designs
  824. 30:35uh for apps, websites, and things like
  825. 30:37paper. Here's what I would sell. I would
  826. 30:40say it's going to cost you 25k up front,
  827. 30:43and it's going to be $3,000 a month.
  828. 30:45But I'm going to include a fully
  829. 30:48redesigned website with that.
  830. 30:50>> Mm.
  831. 30:51>> Um
  832. 30:52I so just completely free.
  833. 30:55So
  834. 30:57I think like and then you show and you
  835. 31:00you have it done for them. Uh you know?
  836. 31:02It's like this is what it's going to
  837. 31:03look like.
  838. 31:05Um and it and it looks beautiful.
  839. 31:08And so they're like, okay, you know,
  840. 31:10we've been thinking about redoing our
  841. 31:11website. Like, yeah, would we pay 25k
  842. 31:13for our website? Yeah. So you're
  843. 31:15basically getting that for free.
  844. 31:17And then you're just paying $3,000 a
  845. 31:19month. And the $3,000 a month, I think
  846. 31:21what's so interesting about this whole
  847. 31:23like model is
  848. 31:24um they're considering it as like a
  849. 31:27hire.
  850. 31:29So like their mental model is $3,000 a
  851. 31:33month. Well, that's, you know, I don't
  852. 31:34know, $12 an hour or something like
  853. 31:36that, you know?
  854. 31:38So I think
  855. 31:40uh that's the other way to think about
  856. 31:41it in terms of pricing is to actually be
  857. 31:4425k up front and to be like $12 an hour.
  858. 31:47>> Yeah. Yeah, that's beautiful. Like
  859. 31:50essentially going back to that concept
  860. 31:53of well, software is so easy to build
  861. 31:56and we can build it so fast, we could
  862. 31:58use it to strengthen our offer and use
  863. 32:01it as a weapon of like, let's just build
  864. 32:03all these assets and just bundle them in
  865. 32:05for free or as bonuses, you know?
  866. 32:08>> Yeah, exactly. The other thing you can
  867. 32:10do to make it irresistible is to say
  868. 32:11like
  869. 32:13you you find out what their bottlenecks
  870. 32:15are, what their pain points are, and you
  871. 32:17could say
  872. 32:19if you aren't able to get back to people
  873. 32:21quicker, if you aren't able to
  874. 32:24you know, text people, you know, via the
  875. 32:28via this, like, you know, you figure out
  876. 32:30those things and you say if you aren't,
  877. 32:32I will give you back
  878. 32:34the 3k a month or whatever or
  879. 32:38like so there's something that's like,
  880. 32:39you know it's no brainer and you know
  881. 32:41you're going to be able to do those
  882. 32:42things because you're you're setting up
  883. 32:44the the Grokbot right?
  884. 32:47So you just you do those so it's like
  885. 32:50okay so you're basically getting a free
  886. 32:51website
  887. 32:54and then you're getting this higher and
  888. 32:56then you're validating that you can
  889. 32:58actually do those things or they get
  890. 33:00their money back.
  891. 33:02>> I love it. I love it. Yeah.
  892. 33:05There's so many there's so many freaking
  893. 33:07like value adds that you can give today
  894. 33:10with AI. Like so many you could build
  895. 33:13them lead magnets for that you could
  896. 33:14build them a a dental you know build a
  897. 33:17calculator lead magnet into their
  898. 33:19website for their dental practice or
  899. 33:21their orthodontist practice to show show
  900. 33:24their clients how much their their
  901. 33:25braces would cost or whatever. Like
  902. 33:27these are like lead magnet lead magnet
  903. 33:29tools you could build them include those
  904. 33:31for free. Like there's so many
  905. 33:34opportunities to just add value today
  906. 33:36for such little cost and time that it's
  907. 33:39like
  908. 33:40selling shouldn't really be the hard
  909. 33:42part here.
  910. 33:44So
  911. 33:46and and so what about this Greg like
  912. 33:49what if I build what if I don't want to
  913. 33:51sell to SMBs
  914. 33:53a custom Grokbot? What if I actually
  915. 33:55want to take a custom Grokbot and sell
  916. 33:57it to an industry? Like maybe I do want
  917. 33:59to build the Grokbot for
  918. 34:02insurance agencies and have it sold you
  919. 34:04know as a one to many tool. Well Grokbot
  920. 34:07they're charging 200 a month
  921. 34:09but if I'm solving this for an industry
  922. 34:11specifically how much pricing power do I
  923. 34:13have?
  924. 34:14What do you think?
  925. 34:17>> I'm going to say it's 10 to 20 x
  926. 34:20>> Yeah.
  927. 34:21>> in terms of like
  928. 34:22you know in general like
  929. 34:25the niching down like that's another
  930. 34:27there's a few reasons why you want to
  931. 34:28niche niche down in general.
  932. 34:31One is you have more pricing power the
  933. 34:33other is you have less competition.
  934. 34:36So I think you you command a higher rate
  935. 34:39for sure.
  936. 34:40>> Got it. Yeah. I agree.
  937. 34:42>> And it's like and you could charge, you
  938. 34:44could do seat-based pricing and have a
  939. 34:46bunch of expansion revenue and like
  940. 34:50and and I just keep going back to this
  941. 34:52point. Like the crazy thing is
  942. 34:54this doesn't need like you don't need
  943. 34:56funding for any of this. Like you can
  944. 34:58just like build this stuff like today,
  945. 35:01you know? Um
  946. 35:03and it's so cool. And so like
  947. 35:05in here I could even tell, you know,
  948. 35:07these bots to like build a website for a
  949. 35:11dental clinic.
  950. 35:12Um
  951. 35:14let me see what what's going on here.
  952. 35:15Hold on. Let me reset this real quick.
  953. 35:20Let's give this a second. There's a bug.
  954. 35:25So in here I could say build a website
  955. 35:28for a dental clinic
  956. 35:30um and load it up please.
  957. 35:33And imagine just having this already
  958. 35:36built into your Grokbot for that
  959. 35:38business
  960. 35:40out of the box. It's like it comes with
  961. 35:42a a website spun up for them. All these
  962. 35:44bots, you know, a a bot that manages the
  963. 35:46website. I actually have one as well
  964. 35:48that I have literally a bot that manages
  965. 35:50my personal website and updates it every
  966. 35:52week. Um
  967. 35:54like if this is just out of the box
  968. 35:56ready for them
  969. 35:57I mean yeah, to me it's like it's a
  970. 36:00no-brainer for a lot of these business
  971. 36:01owners. Um and you can see just how fast
  972. 36:04this thing will build it.
  973. 36:05Um
  974. 36:07while this is cooking, go ahead.
  975. 36:09>> Yeah, it's just one of those things that
  976. 36:11I think that if someone actually wants
  977. 36:12to go and and run this idea like Grokbot
  978. 36:15Grokbot for XYZ I think uh
  979. 36:18in-person sales like you kind of need to
  980. 36:21see it to believe it. You know what I
  981. 36:23mean? Like once you see it, you're like
  982. 36:25oh my god, yes. Like
  983. 36:27why don't I have a someone managing my
  984. 36:29personal website? Why don't I have
  985. 36:30someone managing customer growth? Why
  986. 36:32don't I have someone managing LinkedIn?
  987. 36:34Um
  988. 36:36the you know, it you have to see it.
  989. 36:39>> Yeah, you need the inspiration. People
  990. 36:41don't realize how many things you can
  991. 36:43actually automate. they they they need
  992. 36:45the inspiration in and of itself. And
  993. 36:47like people like us, you know, we're
  994. 36:49we're using these in our daily lives
  995. 36:51>> Yeah.
  996. 36:51>> so much that we have all these ideas
  997. 36:53for, you know, ways that these agents
  998. 36:54can be useful. Um so yeah. Um while
  999. 36:59while this continues to cook. So there's
  1000. 37:01there's those two options there of like
  1001. 37:04if you build your own Grok bot, you sell
  1002. 37:05it to businesses, you could sell it to
  1003. 37:07few and sell for high ticket, or you
  1004. 37:09could sell one to many, still high
  1005. 37:10ticket, uh but just in a more
  1006. 37:12productized approach way. So it's like
  1007. 37:14super scalable. Um and just like
  1008. 37:17recapping on on on the point here of
  1009. 37:19like when you solve the entire gap for a
  1010. 37:21business, it it commands this pricing
  1011. 37:24power, you know? And you and when you
  1012. 37:27add like these guarantees and you just
  1013. 37:29make it irresistible, um that's how
  1014. 37:32you're able to charge high prices. And
  1015. 37:34and if you don't charge this, like a lot
  1016. 37:36of people might have this constraint or
  1017. 37:38like self
  1018. 37:39limiting uh maybe doubt of like, well,
  1019. 37:42if I was able to build this in a
  1020. 37:43weekend, how valuable is that?
  1021. 37:45You don't know how many people Did you
  1022. 37:48see that Andreessen Horowitz post of
  1023. 37:49like 2% of the US pays for AI tools? So
  1024. 37:53if you're building with these things,
  1025. 37:55you're you're taking it for granted how
  1026. 37:57competent of a skill that is and how
  1027. 37:59valuable that is. And um
  1028. 38:01yeah.
  1029. 38:03>> Yeah. Like the vast majority of people
  1030. 38:05haven't heard of Grok bot. They don't
  1031. 38:07know about computer agents. They don't
  1032. 38:08know that you can do this.
  1033. 38:11So
  1034. 38:12and that's going to that 2% is going to
  1035. 38:14go to 4% and 8% and 12 and 18 and 20 and
  1036. 38:1625. So like
  1037. 38:18I agree. I think there is a a limiting
  1038. 38:20belief that some people have around
  1039. 38:23like, well, if someone could create this
  1040. 38:25in a weekend, there's no moat, you know?
  1041. 38:28Um and that's what they want you to
  1042. 38:30think. They want you to think that
  1043. 38:31they're you know You know what the moat
  1044. 38:33is? Like creating value for customers.
  1045. 38:36>> Yeah.
  1046. 38:37>> know what I mean? Like if you create
  1047. 38:38value for for customers, like they're
  1048. 38:40going to stick with you. As long as you
  1049. 38:42create value for customers, they're
  1050. 38:43going to want to stick with you.
  1051. 38:44>> And they want you. Like I think a lot of
  1052. 38:46people undersell themselves. Like,
  1053. 38:49there's something specific and unique
  1054. 38:51about everybody. And And when a business
  1055. 38:55decides to work with you,
  1056. 38:57they're It's not It's not always so much
  1057. 39:00that they're making this decision
  1058. 39:01entirely on the product or service that
  1059. 39:03you're giving them. They're also making
  1060. 39:05it A huge portion of that decision is on
  1061. 39:08you because they like you as a person.
  1062. 39:11They want to work with you long term.
  1063. 39:13And like, so much of this is also like
  1064. 39:15just a relationship thing, you know?
  1065. 39:18Um they want Greg Eisenberg to be the
  1066. 39:20one to build them their their their
  1067. 39:22personal company Grokbot because they
  1068. 39:24like Greg Eisenberg, you know?
  1069. 39:26Um I think people should keep that in
  1070. 39:28mind, too. Like, there's this human
  1071. 39:30element of business that's never going
  1072. 39:31to go away. Um
  1073. 39:33>> Absolutely.
  1074. 39:33>> so
  1075. 39:34yeah. Um so while while that cooks, is
  1076. 39:38there anything else So we went through
  1077. 39:40Okay, what is the the startup idea? The
  1078. 39:42the opportunity. It's like sell
  1079. 39:45Grokbots to businesses, build it
  1080. 39:48customized to their industry. How do you
  1081. 39:50build it? We walked through how to
  1082. 39:51actually build Like, what's the actual
  1083. 39:53strategy when you when you prompt these
  1084. 39:56agents and what is the tool stack that
  1085. 39:57you use?
  1086. 39:59What are all the different components?
  1087. 40:00We talked about that with giving the
  1088. 40:02agent its own email, its own computer,
  1089. 40:04phone number.
  1090. 40:05Um
  1091. 40:07and then finally, you know, talking
  1092. 40:09about the offer. How do we How do we
  1093. 40:11monetize this? What
  1094. 40:13What do we charge? Do we sell one to
  1095. 40:15many, one to few? It depends on the
  1096. 40:17founder type.
  1097. 40:19Um and what's the value proposition?
  1098. 40:21What are some other constraints, Greg?
  1099. 40:23Am I missing anything that people might
  1100. 40:25hold people back from from going and
  1101. 40:27building this opportunity for them for
  1102. 40:28themselves in a weekend?
  1103. 40:30>> There's only one thing, which is
  1104. 40:34like, okay, Nick, like I I see
  1105. 40:37I see the stack. I understand all the
  1106. 40:39all these you know,
  1107. 40:41I see this and But it's like, how do I
  1108. 40:44take
  1109. 40:45this and actually
  1110. 40:48bring this to Willow, well, Willow
  1111. 40:51Dental, to actually
  1112. 40:53like how am I supposed to know which
  1113. 40:55bots I need to create? How do I figure
  1114. 40:57out what the pain points are? Um, what
  1115. 41:00does that look like? What does that
  1116. 41:01process look like?
  1117. 41:02>> Yeah. The So the biggest thing is like
  1118. 41:06deciding your niche is
  1119. 41:08I think a very difficult problem for a
  1120. 41:10lot of people. I And when people get
  1121. 41:13started in business, here's here's a
  1122. 41:15common thing that happens. You'll You'll
  1123. 41:17You'll get super amped up.
  1124. 41:19You'll say, okay, you know what? I'm
  1125. 41:21going to start I'm going to make an
  1126. 41:22agency or I'm going to start selling
  1127. 41:24services or agents or Grok bots to
  1128. 41:26businesses. Okay, and then you go online
  1129. 41:28and you watch videos on on business
  1130. 41:30education and the first thing you're
  1131. 41:32told is pick a niche. And it's like,
  1132. 41:34okay, well, then what's my niche? And
  1133. 41:36for a lot of people, that might be
  1134. 41:37obvious for some cuz they're like, oh, I
  1135. 41:39spent this many years in in dental or I
  1136. 41:42spent this many years in insurance. But
  1137. 41:44for a lot of other folks, they're like,
  1138. 41:45I don't know where to start. I don't
  1139. 41:46know what niche to build. So they you
  1140. 41:48know what the biggest problem that
  1141. 41:49entrepreneurs make is they'll pick a
  1142. 41:51niche.
  1143. 41:53They'll start on that niche. They'll
  1144. 41:54build a website. They'll build all this
  1145. 41:56messaging and offer and positioning and
  1146. 41:57they'll make all this content towards a
  1147. 41:59dedicated niche. And then after a week,
  1148. 42:01they'll realize they don't care about
  1149. 42:03that niche at all. And then they have to
  1150. 42:05start all over again with another niche.
  1151. 42:07So my point here being the first step
  1152. 42:10you should do
  1153. 42:11is don't worry too much about picking an
  1154. 42:14industry or a niche like right away.
  1155. 42:17What I would do is And this is what I
  1156. 42:19did. I'm so speaking from personal
  1157. 42:20experience. I would go through my phone
  1158. 42:23and find the 50 people I know who either
  1159. 42:26are directly like running a business or
  1160. 42:28they're maybe their friends with
  1161. 42:30somebody who runs a business or they're
  1162. 42:32just like high up at a business. And I
  1163. 42:34would just call those 50 people.
  1164. 42:36And I would tell them like, hey,
  1165. 42:39here's the problem I'm solving. I'm
  1166. 42:41building a whole team of agents that can
  1167. 42:43help you run your business and a bunch
  1168. 42:45of apps and like website and everything
  1169. 42:48all in one solution. I'm doing this for
  1170. 42:51businesses. Is this something you think
  1171. 42:53you guys could be interested in?
  1172. 42:55And then,
  1173. 42:56I would, you know, if out of those 50
  1174. 42:58people, you're going to get a few people
  1175. 43:01who are interested. And one of them
  1176. 43:02might be in this industry and one of
  1177. 43:04them might be in that industry. And then
  1178. 43:06I would just do the first one for free
  1179. 43:09just to get a case study of like, okay,
  1180. 43:12asking them a bunch of questions, seeing
  1181. 43:14if we could actually solve these
  1182. 43:15problems for them.
  1183. 43:16And then once you do that and you have
  1184. 43:18that case study, now you might be at a
  1185. 43:20point where, okay, I like this industry.
  1186. 43:23It might be an industry you've never
  1187. 43:24even heard of. Like for me, one of my
  1188. 43:26first one was like a promotional product
  1189. 43:28industry. I didn't even know this
  1190. 43:29industry existed where they make like
  1191. 43:31swag merch for for companies. I was
  1192. 43:34like, wow, really interesting industry,
  1193. 43:35big, untapped it seems. And like I would
  1194. 43:38have never found that if I didn't just
  1195. 43:40be open-minded and just like reach out
  1196. 43:42to the people who were receptive and and
  1197. 43:45just work with that. And so I guess my
  1198. 43:47point being, same way earlier how I said
  1199. 43:49diverge before you converge,
  1200. 43:51same thing is true with like industries
  1201. 43:53and niches.
  1202. 43:54Just try, test out a bunch of them, see
  1203. 43:57what the market appetite is with people
  1204. 44:00that you know and build for them. And
  1205. 44:03then once you're building for them, like
  1206. 44:05just genuinely sit down, don't try and
  1207. 44:08scale right away, sit down with that one
  1208. 44:10first client.
  1209. 44:12Try really to solve their business,
  1210. 44:14watch them, shadow them, ask ask them to
  1211. 44:17like share their screen and go through a
  1212. 44:20process of like something that they do
  1213. 44:22manually, like creating an invoice or,
  1214. 44:24you know, creating products into a
  1215. 44:26presentation and sending that to a
  1216. 44:27customer.
  1217. 44:28And you will see how many specific
  1218. 44:31automations and problems can be solved.
  1219. 44:33And it will be very intuitive, hopefully
  1220. 44:35for you, to build those agents and
  1221. 44:36applications out for them. So.
  1222. 44:38>> Yeah, couldn't agree more. The one thing
  1223. 44:40I'd add is if you are going to go and do
  1224. 44:44a free case study for someone, say a
  1225. 44:46dental clinic, one thing you can do is
  1226. 44:49you can say, hey, I normally charge
  1227. 44:51$50,000 a year for this service.
  1228. 44:53I'm going to do this for you for free.
  1229. 44:57All that I ask is if this works, you can
  1230. 45:00introduce me to three of you know
  1231. 45:03colleagues in in in in the dental clinic
  1232. 45:07space in the local area or the next town
  1233. 45:10over and just give me an intro via text
  1234. 45:13saying like hey you know Nick did this
  1235. 45:16thing for me it it you know it was
  1236. 45:18successful
  1237. 45:20you should consider it.
  1238. 45:21And
  1239. 45:22>> I love that.
  1240. 45:23>> Yeah and and to make it even easier for
  1241. 45:25them you can say
  1242. 45:27uh and I'll give you you know
  1243. 45:30$5,000 credit so it makes them look good
  1244. 45:32or something like that.
  1245. 45:34>> Yeah yeah I'll credit this towards your
  1246. 45:36next month if they're on like some
  1247. 45:38retainer with you or something like
  1248. 45:39that. And and you know the beautiful
  1249. 45:41part about this whole process that you
  1250. 45:43just described too Greg is like at no
  1251. 45:45point ever are you selling cold because
  1252. 45:48you're always if the first person you
  1253. 45:51know is someone you know and you do it
  1254. 45:53for free and then they refer you to
  1255. 45:54someone they know you're always warm and
  1256. 45:57that is the key I think with sales is
  1257. 45:58like you never want to be on a call with
  1258. 46:01somebody who doesn't know you and they
  1259. 46:03don't know who what you sell and it's
  1260. 46:04just cold. It's like it's like dating
  1261. 46:06you know you want to you know I want to
  1262. 46:08meet you through a friend or I want to
  1263. 46:10meet you through some kind of warm
  1264. 46:11mutual interest or connection. That's
  1265. 46:14absolutely true for sales and when you
  1266. 46:16do it like you just described you're
  1267. 46:18always going to be selling warm to
  1268. 46:20somebody uh who has some knowledge about
  1269. 46:22you and what you do. So
  1270. 46:24>> I also think that for for many people
  1271. 46:27including myself like selling is icky
  1272. 46:30like you get kind of this weird feeling
  1273. 46:32but it doesn't need to be that if you're
  1274. 46:33selling something that is so valuable um
  1275. 46:37at a price point that is essentially
  1276. 46:39unheard of that you're teaching them um
  1277. 46:43something that they're really interested
  1278. 46:44in learning about. So there's so many
  1279. 46:46reasons why selling could be this
  1280. 46:49helpful thing that is like they want in
  1281. 46:52so many ways. Um so I wouldn't again
  1282. 46:56like limiting belief like selling isn't
  1283. 46:58icky you know if if you're building
  1284. 47:01something like this and that's why I
  1285. 47:03wanted to bring you on to talk about
  1286. 47:05this because to me this opportunity is
  1287. 47:09a really big one.
  1288. 47:11It's one that you know could could
  1289. 47:14create value you know right away
  1290. 47:18and I think it's one that when you show
  1291. 47:19people the power of this
  1292. 47:22it it really just like you can see their
  1293. 47:24the dots connect in their brain.
  1294. 47:26>> Yeah yeah I think this is like the most
  1295. 47:29exciting time to be building honestly
  1296. 47:32like services or just like any kind of
  1297. 47:36whether you're building a product led
  1298. 47:38business or a service led business an
  1299. 47:40agency or selling agents like whatever
  1300. 47:43kind of business you're building I'm
  1301. 47:46like Greg is rooting for you I'm rooting
  1302. 47:48for you like it's the best time ever to
  1303. 47:50be building
  1304. 47:52and yeah super was super excited to
  1305. 47:54share this startup idea with everyone it
  1306. 47:57keeps me up at night
  1307. 47:59so I hope other people go out and build
  1308. 48:01it too and and the beautiful part is
  1309. 48:03Greg this entire thing that I built if
  1310. 48:06you don't want to take the 36 hours to
  1311. 48:08you know build your own you could fork
  1312. 48:10it it's on GitHub it's you literally go
  1313. 48:12to like
  1314. 48:14the you can go to bops
  1315. 48:16>> We'll include it in the show notes in
  1316. 48:18the description so people can can click
  1317. 48:20it.
  1318. 48:20>> Yeah go to open source tell it point
  1319. 48:23your coding agent to this repo fork it
  1320. 48:25customize it for yourself
  1321. 48:27go sell it go make a lot of money good
  1322. 48:29luck have fun.
  1323. 48:30>> And let us know.
  1324. 48:32>> Yeah and let us know.
  1325. 48:33>> Yeah.
  1326. 48:34>> Yeah and this is the startup ideas
  1327. 48:36podcast if you come up with you know a
  1328. 48:38better twist on it [clears throat]
  1329. 48:40share that too and yeah I think this is
  1330. 48:44just so fun so thanks for having me on
  1331. 48:46Greg.
  1332. 48:47>> Nick thank you for being generous with
  1333. 48:48your time your sauce your ideas your
  1334. 48:51tactics your stack
  1335. 48:53and I'm happy you open sourced this
  1336. 48:56I could speak to you for hours I would
  1337. 48:58love to have you back on. So,
  1338. 49:00you know, comment below what you'd what
  1339. 49:03you want Nick to come back on to to talk
  1340. 49:05about or just anything in general, let
  1341. 49:07us know. Um
  1342. 49:09and
  1343. 49:10uh I'll include links for where to
  1344. 49:11follow Nick on the internet.
  1345. 49:13Um and uh
  1346. 49:16yeah, hope to see you soon, my friend.
  1347. 49:17>> Thank you so much, Greg. Appreciate
  1348. 49:19appreciate you always having me on and
  1349. 49:21um yeah, love love jamming with you and
  1350. 49:24the audience. So, guys, go out there,
  1351. 49:26build some cool things, and um we'll see
  1352. 49:29you next time.
  1353. 49:30>> See you next time.

About this transcript

This page contains the full transcript of The $1M+ Solo Grok Bot Business (Full Course) by Greg Isenberg, generated from the public captions YouTube serves with the video. The transcript has 9,018 words across 1,353 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.