Loading page…
$23K from 4 deals: post build videos on YouTube and let clients come to you
"If your close rate is over 50%, the market is telling you you're charging too little" — across four deals, pricing evolved from a guessed $1,650 to a value-priced $12,000.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
[Video] Started as a solo freelancer: purely out of interest, he posted videos of himself building AI agents on YouTube, with an email address in the description. The first client saw a video and emailed asking for custom work — "I hadn't positioned myself as a freelancer for hire at all." He later partnered with Milan (CEO) and Tyler (CTO) to found True Horizon AI, then added developers and an account manager — roughly a year in total. The four deals: 1) a personalized outreach agent, $1,650 (a guessed quote he now admits was too low); 2) a sales quoting agent, $4,000 (they forgot to capture baseline data, so no ROI case study was possible); 3) a Slack personal-assistant agent, $6,000 (priced by build complexity, which he now calls a mistake); 4) an AI concierge, $12,000 (value pricing + three-tier anchoring: a $25K Scale tier anchors high so $12K looks reasonable).
Nobody was prospected. He was posting videos of himself building AI agents on YouTube because he was excited about learning, with an email address sitting in the description, and a business owner who had watched one emailed asking for the same thing built for him. A couple of calls later the scope was agreed and the deal was closed. His own framing: "at this stage I wasn't positioning myself as a freelancer or saying that I did paid work." That was the entire acquisition engine behind the first four deals — build in public, leave a contact address, answer what comes in. It works because the published artifact is the proof. A recorded build of a working outreach agent is portfolio, technical credential and advertisement in one file, and it reaches precisely the operator who has already decided automation is something he wants. For a shop billing $1,650 to $12,000 a project, this is the cheapest pipeline on offer: publish one build a week showing a named business process automated end to end, and treat every reply as a scoping call rather than a lead to nurture.
The revenue is small, but "post build videos → inbound clients" is a validated zero-cost acquisition path — screen-record your pipeline work and publish it; it doubles as client-getting advertising. The pricing retro section is a free B2B sales course.
Paste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself.
Data credibility: Founder's verbal account — numbers precise to the dollar, but no contracts or payment receipts. All revenue figures are founder-reported or estimated by the video creator; none are audited.
The first invoice was $1,650 for a personalised outreach agent. The client dropped in a contact list, the system researched each person and their company, then wrote a customised opening message and a follow-up into a database, ready to be pasted into an email or DM sequence. It never sent anything. That was the whole scope. The price came from nothing but a previous sale. He had already sold an agent for $1,200, so $1,650 was that number nudged upward because this build felt slightly bigger. He is candid about it: "I was clearly a beginner, and I wasn't even thinking about ROI back then." Run the arithmetic afterwards and the client got a bargain. Two to three hours a week of manual research and writing, valued at $50 an hour, is roughly $400 a month, so the build repaid itself in about four months and saved close to $5,000 over a year. For a first deal that shape holds up: priced off a comparable, scoped to one process, billed as a project rather than a retainer.
Two things compounded here, and neither cost money. The first was the channel. Each build he published brought more inbound, enough that he quit his job and brought in Milan and Tyler to form True Horizon AI, because one person could not scope, sell and deliver at the same time. The second was pricing, which improved faster than the team did. Deal one was a guessed $1,650. Deal two, at $4,000, was argued with time-saved arithmetic: hours saved times hourly rate times four weeks times twelve. Deal three, $6,000, was priced on build complexity, which he now calls the wrong basis. Deal four was value-priced at $12,000, with a $25,000 Scale tier sitting above it so the middle option reads as the sensible one. The inflection was internal systems catching up to deal size — SOPs, an account manager on client communication, a CTO managing engineers instead of connecting nodes himself. That is what let the same shop charge double for the next build. He also names the ratchet: a close rate above 50% means the price is too low, and 20 to 30% is strong for B2B consulting.
The quiet stretch came before anyone paid, and he got through it almost by accident. He was publishing builds on YouTube because he was excited about learning, with no offer, no positioning and no pipeline, so there was nothing to be discouraged by — the work paid him in skill, and the first client arrived on top of that. It is the most durable version of the unpaid period: choose a publishing habit you would keep even if it never sold a thing. The second flat stretch was quieter and more expensive. On deal two they forgot to collect baseline data, which made an ROI case study impossible afterwards — and that case study is the asset that converts one delivered project into the next three. Inbound was arriving faster than their SOPs could absorb, so the CEO delivered the build himself, and on the following project the CTO was personally wiring workflow nodes. The readings that matter in that window are unglamorous: how many inbound emails become scoping calls, what share of proposals close, and whether the finished build produced a number you can show the next buyer.
[Video] The process: a discovery call to surface needs → bring in the CTO to scope the technical range → a second discovery call → the CEO closes. The pitch sells outcomes, not tools: don't say "I can build a chatbot" — say "your team spends 15 hours a week answering repeat questions; I can cut that to nearly zero." Key retro: their early close rate topped 50% — "that's the market telling you you're charging too little" (healthy B2B consulting closes at 20-30%). Pricing evolution: gut guess → time-saved ROI (hours saved × hourly rate × 4 weeks × 12 months) → value pricing + tiered anchoring. Against small early retainers: "one $10,000 project beats five $2,000/mo retainers" — stability comes from a lead pipeline, not monthly contracts.
[Video] A deliberately light stack: n8n + a vector database + 1-2 AI models; deal #4 introduced the then-newly-hot MCP servers, using "staying on the frontier" as a price-raising lever. QA process: 1 week internal (stress-testing with real data + edge cases) → 1 week of client testing and feedback. Lessons: the CTO ended up personally wiring nodes — team-role confusion; and saying yes to "just one more feature" without repricing eats your margin. [AI] Extrapolation: most so-called AI agents are workflow automation plus LLM nodes — the hard part is mapping the client's business process, not the tech. That's exactly why the barrier is low for indie developers.
[AI] Get the scale right: $23K is total gross, split three ways, all clients anonymous and unverifiable — and the video sits at the top of the funnel for his School community and "Automation to Monetization" course, while the "as a beginner" title misleads (the last 3 deals were done as a company). But "post build videos → inbound clients" is a genuine zero-cost acquisition path — as you build your automated site pipeline, screen-record every step and publish it; the footage is itself client-getting advertising. If someone shows up, close them with his pricing ladder; if nobody does, you've lost nothing.
That was one of the 50 breakdowns open to everyone. The Unicorn Club opens the other 1,285 at the same depth — plus pain-point filters and 10 build specs a week.
Join the Unicorn Club — $5/moCancel anytime, in one click