From 47 Ideas in My Notes App to $1,200 MRR in 60 Days — How an AI Co-Founder Killed 44 of Them

On the first of January this year, a solo founder we'll call Marcus had forty-seven startup ideas saved in a Notion document titled "Ideas - Do Not Delete." He'd been adding to it for four years. The ideas ranged from a subscription service for plant care advice to a B2B tool for logistics companies to manage driver compliance documentation. Some had one line. Some had three pages of notes. None had been tested, built, or taken further than the document.

By the first of March — sixty days later — Marcus was generating $1,200 in monthly recurring revenue from one of those ideas. The other forty-four were dead. Two were still alive but deprioritised. The path from forty-seven undeveloped ideas to one working business in sixty days is worth examining in detail, because the process that produced it is repeatable.

The first week was spent on triage. Marcus used an AI co-founder — specifically, Kooio — to run each idea through a rapid evaluation against four criteria: Is the problem specific? Is the target customer identifiable and reachable? Is there evidence of existing demand somewhere — in forums, in Reddit threads, in the way people currently spend money? And does the founder have any genuine insight or access that a stranger wouldn't have? Ideas that failed two or more criteria were cut immediately. That eliminated twenty-nine ideas in four days.

The eighteen survivors were then subjected to a second round of evaluation focused on competitive dynamics. For each idea, Marcus ran a quick market map: who already plays in this space, how well-funded are they, and what would have to be true for a bootstrapped solo founder to carve out a durable niche? Seven more ideas were cut at this stage — not because the market was too small, but because the incumbents were too entrenched and the differentiation Marcus could plausibly offer was too thin.

The eleven remaining ideas moved into a one-week rapid research sprint. Marcus spent five to six hours on each: reading forums, talking to three to five potential users informally, and assembling a one-page assumption map. The goal was not to confirm the ideas but to find the ones that generated the most evidence of real pain with the clearest demand signal. By the end of this sprint, the field had narrowed to three ideas.

The three surviving ideas were: a documentation tool for freelance compliance consultants, a client onboarding kit generator for small agencies, and a driver briefing document builder for small fleet operators. All three had the same profile: specific problem, identifiable customers, no dominant tool, and domains where Marcus had adjacent knowledge from a previous career in logistics and professional services consulting.

Marcus ran a demand test on all three simultaneously. For each idea, he built a simple landing page in a day using a no-code tool, wrote a clear value proposition, and ran $150 of targeted ads to the relevant audience. He set a threshold in advance: any idea that generated ten or more sign-ups in one week would move forward. Only two ideas hit the threshold. The driver briefing document builder generated thirty-one sign-ups in six days. The agency onboarding tool generated fourteen. The compliance documentation tool generated four.

The driver briefing document builder was the clear winner. Marcus killed the compliance idea. He parked the agency onboarding tool — the demand was real but the conversion rate was lower and user interviews revealed more friction than the briefing tool. He now had one idea with a validated demand signal, a clear target customer, and thirty-one people who had raised their hand.

Over the next three weeks, Marcus built the first version — a web app that takes a route description and driver profile and generates a structured pre-trip briefing document. He used AI to accelerate the build: Cursor for code, GPT-4 for document generation, Vercel for hosting. The total build cost was under three hundred pounds. He shared the product with the thirty-one sign-ups, offering three months free in exchange for feedback. Twenty-two activated.

Charging started in week eight. Marcus offered a simple flat-rate plan at fifty pounds per month. Twenty-four of the twenty-two active users converted to paid. Two churned immediately. By the end of week nine, he had twenty-two paying customers and $1,200 in monthly recurring revenue.

The lesson Marcus drew from this is not that the process is easy. It is that the process is clarifying. Having forty-seven ideas is not a disadvantage — it is a resource. But without a structured method for killing ideas quickly, the list never becomes a business. The AI co-founder function here wasn't building the product or writing the copy. It was providing the framework for rapid, unsentimental evaluation — forcing criteria-based decisions at each stage rather than letting intuition and attachment delay them.

The most important number in this story is not $1,200. It is forty-four — the ideas that got killed. Those forty-four were not wasted. They were the work. Every founder with a notes app full of ideas has the same starting material. The question is whether they have a process for turning it into signal or whether the list just keeps growing.