GenAIPI
What is not verified
The headline, a US$1 million a year run rate in September 2025 while the founder worked alone, rests on one statement by the founder, Jon Cheney, on the Mixergy podcast on 23 April 2026, seven months later; the company's own blog gives a lower figure for that month. The later figures, US$2.5 million in the first year and US$4.5 million a year in April 2026, come from the founder too, the first through a podcast description and the second from the same interview, repeated in the interviewer's newsletter on 1 May 2026. No payment processor, accountant, buyer or journalist has seen the books, and the US$20 million valuation is the founder's own. A reader could look for a payment-processor profile, a second interview with the same figures, or client confirmations of the contract sizes. The headline number is grade D: one claim or one estimate, with no second source. Aiolum shows it because the story is useful, and marks it so that it is not mistaken for a verified figure.
GenAIPI, the General AI Proficiency Institute, is a US consultancy that installs AI inside other companies. Its main service is a fractional Chief AI Officer: a senior person who acts as a company's head of AI for a few days a month instead of being hired full time. The engagement includes a 90-day plan, hands-on building of workflows with the client's staff, role-specific training, and since 2026 a software product, G.E.N.A., which the founder describes as an agent harness, a layer that lets AI agents read a company's systems and carry out tasks with permission checks. The founder said in April 2026 that engagements are priced from US$10,000 a month, and the company's own page says the same.
Jon Cheney, who said he raised US$13 million for an earlier venture-backed startup founded in 2016 and later sold it, built the first version by vibe coding, which means describing an application in plain language to an AI coding tool, here Replit, and letting it write the code. He told the Mixergy podcast in April 2026 that he started on Thursday 27 February 2025, launched on 1 March with US$400, found that his first Facebook ads produced 140 visitors and zero sales because Stripe was misconfigured, and sold a US$15,000 training contract to a local CEO through LinkedIn messages on the Monday. He said in the same interview that he reached a US$1 million a year run rate, the current month's revenue times twelve, in September 2025 while still working alone, with six Chief AI Officer clients at once; that figure is the headline of this page, because it is the last one from the time he was the whole company. He then hired. A podcast description reposted in March 2026 said the company had made US$2.5 million in its first year with "a handful of employees", and he said in April 2026 that it was at about US$4.5 million in annual recurring revenue, the yearly value of the retainers in force, with a team of unstated size.
Every number on this page is a statement by the founder, mostly from one interview published twice, so the page carries grade D, reported but not verified; the headline run rate was stated seven months after the fact. No payment processor, buyer or journalist has checked the figures, and the company's own blog gives a lower figure for month six than the founder gave on the podcast. The page also shows what happened to earlier companies that sold AI or training to other businesses.
The numbers, graded
Each row shows the evidence grade and the source numbers from the list at the end of this page. The grades are explained in How Aiolum verifies.
| Fact | Value | Grade | How we know | Sources |
|---|---|---|---|---|
| Revenue run rate, September 2025, founder working alone | US$1 million a year | D | The headline figure. Founder on Mixergy, 23 April 2026: "I hit a million in September", "about 80,000, 90,000 a month", later "110,000 a month", "by myself". The company's blog gives "$50K+/month" for month six, which is lower. | [1, 4] |
| Annual recurring revenue, April 2026, with a team | about US$4.5 million a year | D | Founder on the Mixergy podcast, 23 April 2026: "We're sitting at about four and a half million in annual recurring." The interviewer's newsletter repeated the figure on 1 May 2026. One interview, two publications, so one source. Headcount at this figure not stated, so it is not the headline. | [1, 2] |
| Revenue in the first year, March 2026 | US$2.5 million, over 50% profit | D | Description of The Koerner Office podcast episode, as reposted by Ventureboom in March 2026: "A year later he's at $2.5 million in revenue, over 50% profit, and on pace for $7 million this year with just a handful of employees." | [3] |
| Total revenue in the first six weeks, April 2025 | over US$100,000 | D | Company blog post, shown as dated 10 May 2025: "it generated over six figures in revenue within the first 6 weeks". | [4] |
| Monthly revenue in month six, September 2025 | US$50,000 or more a month | D | The same blog post: "Month 6: Processing $50K+/month consistently". The post's shown date, 10 May 2025, is earlier than the sixth month it describes, so the post was changed later or the date is wrong. | [4] |
| First contract, 3 March 2025 | US$15,000 | C | Founder on Mixergy, the Replit customer story and the company's press page agree: a one-day AI training for a local company's staff, sold by LinkedIn message on the Monday after launch. | [1, 5, 6] |
| First fractional Chief AI Officer contract, 2025 | US$15,000 a month, US$180,000 a year | D | Founder on Mixergy: invented on a call with a pool company owner who asked for ongoing help; signed by DocuSign and paid the same day. | [1] |
| Starting cost | US$400 | C | Founder on Mixergy and on the Replit customer story. The blog puts running costs at under US$100 a month: Replit US$25, database under US$20, SendGrid US$10. | [1, 4, 5] |
| Profit margin | about 99% at the start, about 50% in April 2026 | D | Founder on Mixergy. Clients paid tool costs on their own cards; the margin fell when he hired people and added software. | [1] |
| Valuation, 18 months after launch | US$20 million | D | Replit's customer story: "$20M valuation after eighteen months, according to Jon". No transaction or investor is named; it is the founder's own number. | [5] |
| Website visits in one week, April 2026 | 876 from organic search, 412 direct, 166 organic video, 105 organic social, 44 referral | D | Founder reading his Google Analytics on Mixergy. He said the GEO work made ChatGPT recommend the firm to a large training company. | [1] |
| Team at the US$1 million run rate, September 2025 | 1 person with six officer clients at once | D | Founder on Mixergy: "I took this to a million dollars by myself"; "I had, you know, six, you know, CAIO jobs at the same time"; then he started hiring. | [1] |
| Team, March to April 2026 | "a handful of employees"; headcount not stated | D | Podcast description reposted by Ventureboom, March 2026. Founder on Mixergy: "I've got a team now that's working and delivering." The site also runs a Talent Network of operators who can be "deployed into companies". | [3, 1, 16] |
| Founder's earlier company | raised US$13 million, founded 2016, later sold | D | Founder on Mixergy. The company's about page says "$20M+ raised" and "Multiple exits" over 20 years. | [1, 10] |
| Outside funding for GenAIPI | none stated | D | The founder describes starting with US$400 and says he did not want to raise money again. No round is reported anywhere. | [1, 2] |
| Projection stated in April 2026 | US$10 million a year; US$1 million a month by December 2026 | D | Founder on Mixergy. A forecast, not a result. | [1] |
| Time from launch to the September 2025 figure | 6 months | calc | Calculated: March 2025 to September 2025. The April 2026 figure came 13 months after launch. | [1, 4] |
How to read this table: a grade D value is a founder statement made on one occasion, with no second source that agrees on the same period. A grade C value is a founder statement that two sources, or the company's own pages, give alike. A "calc" value was computed on this page from the rows above it. No value is grade A or B.
| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Fractional Chief AI Officer, from the AI Leadership System level | from 10,000 | companies with 10+ employees | |
| Live course: How to Build a Business in a Weekend with AI | 1,349 one-time | one day, taught by the founder | |
| Live course: How to Build a Business with AI, Setting Up OpenClaw | 345 one-time | one day | |
| AI Proficiency Certification Test | 59 one-time | 40 questions, 60 minutes, 70% to pass | |
| AI Readiness Assessment and AI Mastery Predictor | 0 | free, email address required |
The Chief AI Officer engagements have four levels (AI Leadership System, AI Enablement, AI Transformation, Enterprise Transformation); the page gives only the starting price and says the level is set on a call. The page says the service is for companies with 10 or more employees. The courses are one-day live sessions. G.E.N.A. has no public price; the founder said in April 2026 that it is sold together with the officer service, not alone.
The timeline
From a US$105,000 app quote to a one-person US$1 million run rate in six months, then a team and a consultancy the founder values at US$20 million, all in his own words. Each entry carries its grade.
- 2016DCheney founds a venture-backed startup and, he says, raises US$13 million over the years; the company is later sold. [1]
- Late 2024CA development agency quotes US$105,000 for a travel app he has wanted for years. The same day he reads a LinkedIn post about Replit, drops the agency's proposal PDF into it, and has a working first screen in ten minutes. Over the next months he builds about 20 apps. [1, 2, 5]
- 27 February 2025DHe starts building the General AI Proficiency Institute: a free AI test that collects an email address, then a paid certification and courses. Three days of about 12 hours each, with the LLC, bank account and social pages set up while Replit works. [1]
- 1 March 2025CLaunch. Facebook ads go live at 3 a.m. on Saturday night. By Monday, 140 people have started the test and nobody has paid: Stripe was misconfigured. [1, 4]
- 3 March 2025CHe messages people on LinkedIn, gets a call with a local CEO and ends it with a US$15,000 contract to test and train the company's staff. [1, 5, 6]
- 27 March 2025CThe Wayback Machine holds a first snapshot of genaipi.org. [14]
- April 2025DThe company's blog says revenue passed US$100,000 in the first six weeks. [4]
- 7 May 2025CThe local station KUTV runs a story: AI helped a Utah entrepreneur build a six-figure business in a few days. The founder says he called the station himself. [6, 1]
- 10 June 2025DPress release on PR Newswire: the institute has delivered a US$75,000 enterprise training. [6]
- Mid-2025DA pool company owner asks what he offers after the training day. He invents the fractional Chief AI Officer on the call at US$15,000 a month and sends a DocuSign contract the same day. [1]
- September 2025DThe founder says he hit a US$1 million a year run rate, alone, with six officer clients, and started hiring. [1]
- March 2026DA podcast description says US$2.5 million of revenue in the first year, over 50% profit, and "a handful of employees". [3]
- 23 April 2026DOn Mixergy he says the company is at about US$4.5 million in annual recurring revenue and announces G.E.N.A., a company agent built on NVIDIA's NemoClaw, to launch in about ten days. [1]
- 1 May 2026DThe interviewer's newsletter repeats the US$4.5 million a year figure and the playbook. [2]
- September 2026DReplit's customer story says GenAIPI was worth US$20 million eighteen months after launch, according to the founder, with employees and health insurance. [5]
- 24 October 2026CThe site sells officer engagements from US$10,000 a month, a US$1,349 one-day course, a US$345 OpenClaw course and a US$59 certification; the next course session is 17 November 2026. [7, 8, 9]
Why it worked
Four causes, in the order they mattered. Each one is a reusable lesson.
A credible shell in a weekend
The founder said in April 2026 that the name, the .org domain, the social pages, the bank account and the working payment flow were all set up before the first ad, so that a buyer would find no broken link. He says the institutional look is what let him sell to governments and larger companies at a price of US$15,000 a day.
Sell the service, not the course
The automated plan, free test then paid certificate, made nothing in its first weekend. The money came from a direct LinkedIn message and a call. The founder told Mixergy that his advice is to skip the ads and message customers on LinkedIn, then look for recurring revenue.
Turn the training day into a retainer
A customer asked for ongoing help; the answer was a part-time head of AI priced like an employee, which the founder said in April 2026 was US$10,000 to US$15,000 a month, cancellable at any time. He says the first such contract was worth US$180,000 a year and that he will not take clients under US$20,000 now.
Let the client pay for the tools, then add software
The founder said in April 2026 that clients paid Replit and other tool charges on their own company cards, so his margin was close to 99% while he worked alone. In 2026 the firm packaged what it built for clients into G.E.N.A., an agent product it installs together with the officer service, because, in his words, software alone is almost worthless now and distribution is the moat.
How it was built
These details come from the founder's podcast interview, the company's blog and Replit's customer story. They are grade C or D. [1, 4, 5, 2]
- The site, the admin tools, the tests and the course platform were built in Replit by a founder who says he does not know how to code; the blog says Replit also hosts it, with a database that scales with use and SendGrid for email.
- Stripe takes payments for courses and certification; the first launch failed because it was misconfigured.
- ChatGPT was used to choose the business, by interviewing the founder about his skills, and Grok to test the plan against what people on X say.
- Search visibility: he had ChatGPT and Grok study a leading generative-engine-optimization firm, merged their findings and told Replit to install the result, meta tags, robots file, FAQs; he said in April 2026 that half of visits came from organic search and that a large training company found the firm through ChatGPT.
- Sales come from daily LinkedIn posts, direct LinkedIn messages, local television and podcasts; DocuSign for contracts.
- The agent product G.E.N.A. is a customized NemoClaw, NVIDIA's enterprise harness for the open-source OpenClaw agent, connected to each client's systems such as NetSuite, QuickBooks, HubSpot or Salesforce over a few weeks.
The dead that came before
Four companies in the Loot Drop database sold AI, automation or skills to other businesses and failed. The values below are from that database and are not independently verified.
Olive AI
The promiseAn AI workforce that would automate hospital administration.
Why it diedEach client needed custom integrations built by engineers, the bots broke when systems changed, and the cost of delivery never matched the price of the promise.
Pathmind
The promiseLet companies use AI to optimize operations without in-house experts.
Why it diedCustomers were not ready, the product was complex, and the big cloud providers offered the same capabilities with more reach.
Pluralsight
The promiseKeep company workforces skilled through on-demand technology courses.
Why it diedExpensive enterprise sales for seats that few employees used, while free video and AI made the content less valuable.
Zirtual
The promiseA subscription assistant for busy business owners.
Why it diedA people-heavy service grew faster than its finances; it closed overnight when it could not meet payroll.
What changed
Olive AI and Pathmind sold automation to companies whose systems were all different, and the cost of making it work at each client ate the business. Pluralsight sold training seats that employees did not use. Zirtual sold a people-based service by subscription and could not pay its people. GenAIPI combines all four promises: automation, implementation, training and a retainer for a person's time.
What differs, in the founder's account, is the cost side. He says the first US$1 million a year of revenue was earned by one person whose tools cost under US$100 a month, with clients paying for their own software, so the margin was near total and no money was raised. The custom integration work that sank Olive AI is now done with AI coding tools at the client's expense, and the retainer is priced like an employee rather than like a software seat. Whether that holds as the team grows is the open question; he said in April 2026 that margins had already fallen to about 50%.
Can one person copy it now?
The model, yes. The numbers, not verified.
The founder says anyone can copy him and that the window is about five more years. The parts that do not depend on his background are concrete: a credible site built in days, direct outreach instead of ads, a training day that becomes a monthly retainer, and the client's own card for tool costs. The part that is hard to copy is a decade as a funded CEO, which is what let him charge US$15,000 a day from the first call, as he said in April 2026.
Where it is still open: Every trade with many similar small companies, law, dental, logistics, construction, that has no in-house AI lead and can pay US$10,000 a month; and the agent-harness layer, OpenClaw and NemoClaw, that lets a consultancy sell software it did not write.
- 1Build the complete storefront, domain, payments, social pages, in a weekend with an AI coding tool, and test the payment flow before any ad runs.
- 2Message the people you already know on LinkedIn and sell a fixed-price training day; do not wait for the automated funnel.
- 3When a client asks what comes next, offer a part-time role priced like an employee, cancellable monthly, and sign it the same day.
- 4Put tool and model costs on the client's card, keep your own costs under US$100 a month, and package what you build for one client into a product for the next.
What to watch
The risks are the mirror image of the causes above.
What is not verified
The headline, a US$1 million a year run rate in September 2025 while the founder worked alone, rests on one statement by the founder, Jon Cheney, on the Mixergy podcast on 23 April 2026, seven months later; the company's own blog gives a lower figure for that month. The later figures, US$2.5 million in the first year and US$4.5 million a year in April 2026, come from the founder too, the first through a podcast description and the second from the same interview, repeated in the interviewer's newsletter on 1 May 2026. No payment processor, accountant, buyer or journalist has seen the books, and the US$20 million valuation is the founder's own. A reader could look for a payment-processor profile, a second interview with the same figures, or client confirmations of the contract sizes.
The company's own blog gives a lower figure
The blog post about building on Replit says month six was "$50K+/month", about US$600,000 a year, while the founder told Mixergy that month six, September 2025, was US$80,000 to US$110,000 a month. The blog may count only payments through the platform, or it may have been edited later: its shown date, 10 May 2025, is before the month it describes.
Team size after the headline figure is unknown
The founder says he reached the first US$1 million a year alone, which is why that figure is the headline. By March 2026 there were "a handful of employees", by April a team "working and delivering", and the site recruits a Talent Network of operators to be deployed into client companies. If the firm is larger than ten people now, it is a small consultancy, not a one-person AI business.
A consultancy, with a product attached
Most of the revenue is retainers for people's time; the founder himself calls that a linear business. G.E.N.A., the agent product, is weeks old at the time of the interview and is sold only together with the officer service.
The courses sell the founder's own story
The US$1,349 course is called How to Build a Business in a Weekend with AI and promises "a fully functional, revenue-generating business in just one weekend" and "from idea to paying customers in days". No income figure is promised to buyers, but the marketing is the founder's result. The founder told Mixergy that the original plan, to sell certificates and courses automatically, "was a bust".
Figures that are forecasts
"On pace for US$7 million this year" (March 2026) and "US$1 million a month by December" (April 2026) are projections stated by the founder, not results.
Sources
The grade next to each source is the best grade that the source can support. "Read in full" means that the text of the page was downloaded and the numbers were checked in it.
- D Mixergy, 23 April 2026: "I make $4.5 million implementing AI", interview with Jon Cheney by Andrew Warner, with full transcriptRead the full transcript. Source for the US$4.5 million ARR, the US$1 million run rate in September 2025, the launch dates, the first contracts, the margins, the team statements and the tools.
- D The Next New Thing newsletter, 1 May 2026: He makes $4.5 mil/year implementing AI, this is his playbook, by Andrew WarnerRead in full. The same interviewer's summary of the Mixergy interview; the US$4.5 million a year figure, the US$400 start, the Stripe failure, the daily LinkedIn posts and the NemoClaw move.
- D Ventureboom, March 2026: He Turned $400 Into $2.5M Using AI (No Coding), a repost of The Koerner Office podcast episode descriptionRead in full. A repost of a podcast description, used only for the first-year revenue, profit and headcount sentence.
- D GenAIPI blog, shown as dated 10 May 2025: How I Built a Multi-Million Dollar Business with Replit for a Few Hundred Bucks, by Jon CheneyRead in full. Source for the 1 March 2025 launch date, the six-week and month-six figures, the running costs and the tool list. Its shown date is earlier than the sixth month it describes.
- D Replit customer story: GenAIPI, Built in a Weekend for Under $400Read in full. A vendor page quoting the founder: US$400, US$15,000 first customer within five days, US$20 million valuation 'according to Jon'. Its page metadata carries a date before the company existed, so it is cited undated.
- C GenAIPI press page: KUTV story of 7 May 2025 and press release of 10 June 2025Read in full. The company's own list of its press, with the dates and summaries of the television story and the press release.
- C GenAIPI: Fractional Chief AI Officer page, read 24 October 2026Read in full. Starting price of US$10,000 a month, the four engagement levels, the client list and the 10-employee minimum.
- C GenAIPI: Live AI Training Courses page, read 24 October 2026Read in full, prices from the page's structured data. The two one-day courses, their prices and the next session date.
- C GenAIPI: Certification page, read 24 October 2026Read in full. The US$59 test, 40 questions, 60 minutes, 70% to pass; the free assessments.
- C GenAIPI: About pageRead in full. Founded in early 2025 by Jon Cheney; the founder's 20 years, US$20 million raised and multiple exits before it.
- C GenAIPI home page, read 24 October 2026Read in full. The five-part offer, the statement that the founder leads a multidisciplinary team, and the list of media appearances.
- C GenAIPI: G.E.N.A. product pageRead in full. The agent product's description as an organizational agent harness with permissions; no public price.
- C GenAIPI: Success Stories pageRead in full. Client case studies and testimonials; used for the kind of work delivered, not for revenue.
- C Wayback Machine snapshot of genaipi.org, 27 March 2025Opened; the archived page renders by script and shows only the company name, which confirms the site existed on that date.
- C Loot Drop database export, 1,749 failed startupsLocal export. AI-written summaries, not independently verified. Source for the failed predecessors.
- C GenAIPI: Talent Network pageRead in full. The company recruits and certifies operators who may be deployed into client companies.