With six months of runway, choose the pilot that puts your product inside a painful, recurring workflow and creates a clear path to paid use. Visibility from a celebrated tech hub can help later, but it rarely replaces evidence that someone will change how they work and pay to keep the result.
Friday made the trade-off visible
At 4:40 on Friday afternoon, Mina was sitting in a café in Osu with her laptop on 12 percent battery and two calendar holds open beside each other.
The first was a call with a startup programme in Berlin. They wanted her AI operations tool in their next founder cohort. The offer came with introductions, a demo slot, and the kind of logo that would make future investors pause on a slide.
The second was a visit to a food manufacturer outside Accra. Its operations manager had sent Mina three photos of handwritten production notes, each covered in corrections. He wanted to know if her tool could turn the notes into a usable record before the next shift change.
Mina had six months before payroll and cloud costs forced a harder conversation. She could not build deeply for both.
The Berlin pilot offered a room full of people who might talk about the product. The manufacturer offered one team with a problem that showed up every day, but it was less polished. There would be no launch post, no audience of investors, and no guarantee that a factory manager would become a paying customer after the test.
If Mina chose badly, she could spend the next eight weeks preparing impressive demos while her runway disappeared. The manufacturer might also decide that paper was cheaper than changing a process that had worked for years.
Visibility can hide a weak learning loop
A celebrated hub can create useful momentum. It can sharpen a pitch, widen a network, and give a young company a credible room to enter. Those are real benefits, especially for founders building across African, European, and US markets where the next customer or collaborator may be in another time zone.
But a visibility pilot often has a weak learning loop.
The people watching a demo may not own the workflow. They may be curious about AI without carrying the cost of an error, a late handover, or a missing record. Their feedback can sound enthusiastic because it costs them nothing to be enthusiastic.
Mina looked again at the Berlin brief. It asked for a product walk-through, office hours for founders, and a short presentation on how AI could support small teams. None of it required a participant to replace an existing process. None of it identified who would approve a budget if the cohort wanted to continue.
That did not make the opportunity bad. It made it a distribution activity, and Mina’s immediate constraint was product learning tied to revenue.
A pilot earns priority when it answers a decision that changes the company’s next move: whether to build a feature, narrow a market, price a workflow, or stop pursuing an assumption. The same distinction appeared in Nia’s approval gap: a pilot becomes fragile when the people using a tool cannot safely approve what it does.
The manufacturer gave Mina a harder, better test
On Monday morning, Mina stood beside the manufacturer’s operations manager as a supervisor flipped through the previous day’s notes. A page had a coffee ring near the corner. Another had a quantity crossed out twice.
The manager did not ask for an AI strategy. He asked whether Mina’s product could capture the notes, flag missing fields, and leave a record his team could check before the shift ended.
That request created boundaries.
Mina could define the first workflow in a sentence. She could identify the person who would use it, the person who would judge it, and the error that would make the pilot fail. She could also ask the question that matters before writing more code: if this removes enough rework, who signs off on paying for it?
The answer was incomplete. The manager needed to speak with finance, and the supervisors were wary of entering the same information twice. Mina did not leave with a contract.
She left with a test she could run.
For the next two weeks, she cut the pilot to one production handover. No broad dashboard. No custom analytics. No promise that the tool would solve every reporting issue in the factory. She watched where people paused, which fields they skipped, and which correction came back most often.
That narrowness protected the roadmap. It also gave Mina a more honest result than a room applauding a demo.
Choose the pilot that can change your next Monday
By the end of the test, Mina had learned that the missing-field alert mattered more than the summary screen she had planned to build. The operations manager cared about catching an incomplete record before it travelled to the next person. The supervisors needed a faster way to correct an entry without starting again.
The Berlin programme was still there. Mina asked to join a later cohort, once she could show a workflow that had survived contact with a real team.
That choice did not reject visibility. It put visibility behind evidence.
When two pilots arrive at once, write down the decision each one can answer. Then look for the pilot with a named user, a recurring problem, an owner with authority, and a next step toward payment. If those pieces are missing, treat the opportunity as exposure and price the time accordingly.
Mina’s laptop was charged when she opened the Berlin invitation again. This time, she had a sharper product to bring into the room.
Comments
No comments yet.