The demo should open doors overseas, but the roadmap should follow the workflow local customers are already trying to complete. When attention and usage point in different directions, treat the demo as evidence of interest and the repeated automation request as evidence of demand.
At 4:40 on a Thursday afternoon, Kojo was standing beside a whiteboard in Accra, refreshing a video-call window while a prospect in Berlin asked his team to show the AI assistant again. The demo turned a rough sales note into a clean follow-up and suggested the next action. Everyone on the call leaned forward. One person asked whether the team could support a pilot.
Kojo had spent the previous two weeks preparing for that moment. His cofounder had pushed a late-night build. Their designer had made the interface look calm enough to hide the parts still held together by manual checks.
Then, after the call, Kojo opened the support inbox.
Three local customers had written about the same thing: they wanted the system to send a payment reminder when an invoice was overdue, then mark the account for follow-up if nobody responded. The rule was plain. Their team had already built most of it. It worked in their own test account.
The Berlin prospect wanted a proposal by Monday. The local customers were threatening to return to spreadsheets before month-end. If Kojo gave the engineers to the overseas pilot, the reminder automation would wait. If he shipped the automation first, the prospect could decide the team had moved too slowly and take the conversation elsewhere.
For a small SaaS team, this is one of the harder moments to read correctly. Attention feels like a verdict. It often arrives with better vocabulary, a larger market, and the possibility of a contract that changes how the team is perceived. A repeated operational request can look modest beside it.
But local customers do not ask for an overdue-invoice reminder because they are less ambitious. They ask because someone still has to notice the missed payment, decide what to say, and remember to follow up. The work keeps landing on a person until the product takes it away.
The demo revealed curiosity, the request revealed a recurring job
An AI demo can be useful before it is commercially useful. It can show that buyers understand the category, that a new interface creates interest, or that a team can explain its product beyond its home market. Those are real gains.
They do not answer the more expensive question: what will somebody keep paying to avoid?
Kojo’s three customers were describing a recurring job in different words. One said her finance person kept “chasing the same people.” Another wanted “something that nudges them without us remembering.” The third had already written a version of the rule on paper for his operations lead.
That language matters because it points to ownership. The customer knew exactly where the work began, who carried it, and what happened when it slipped. The AI demo produced admiration. The reminder rule sat inside a process that already had consequences.
This is why workflow ownership becomes more valuable than a clever output. If the product owns a step that people cannot afford to skip, it gains a place in the customer’s week. The question is explored more deeply in AI product defensibility: What Miriam’s Question Taught Founders About Workflow Ownership.
The tempting overseas deal can distort the evidence
A prospect overseas may be a strong signal. It may also be a clean room, far from the constraints shaping the existing product.
The prospect in Berlin had seen a polished sequence in a controlled call. Kojo’s local customers were dealing with actual overdue accounts, staff handoffs, and the awkwardness of sending a second reminder to someone they knew personally. Their request had less theatre. It carried more operational detail.
That distinction should change how a founder allocates a week.
The team can still answer the prospect. They can send a short proposal, describe the current product honestly, and invite a conversation about the workflow behind the demo. What they should avoid is quietly turning one interested call into a new product strategy.
A good test is simple: if the overseas pilot disappeared tomorrow, would the work still make the existing customers more likely to stay, use the product, or expand it? If the answer is no, the work belongs in a separate bet with a defined cost.
Kojo’s team had limited runway. Every new branch created support obligations, edge cases, and meetings that would pull someone from the product customers were already using. The next engineering decision had to account for that cost, not only the possible upside.
Ship the small automation with a clear boundary
On Friday morning, Kojo chose a narrower path. One engineer finished the reminder workflow. The team defined what would trigger it, what message the customer could edit, and when the account would appear for a human follow-up. They did not add AI-generated collections messages or build a new analytics screen around it.
The demo remained available. The proposal went out with a direct note: the team could explore a pilot after understanding the prospect’s payment and follow-up process. That sentence slowed the sales conversation down. It also stopped the team from promising a version of the product they had not decided to support.
By the following week, one of the local customers had changed the reminder text to sound like her own team. The follow-up list gave her operations lead a smaller set of accounts to call. The work had not disappeared, but it had become visible before it became urgent.
That is a stronger product decision than chasing applause. It turns a repeated request into a dependable part of the customer’s process.
Keep the evidence separate until it earns a merge
Founders building across African, European, and US markets often have to hold conflicting signals at once. A foreign prospect may validate a story the local market has not yet rewarded. A local customer may expose a problem that looks too ordinary to impress anyone on a demo call.
Keep both signals. Label them accurately.
The AI demo can be a distribution asset, a way to start conversations and learn which buyers lean in. The working automation can be a retention asset, a reason a customer returns to the product on a difficult Monday.
Kojo’s whiteboard was still covered with ideas for the assistant. He did not erase them. He moved the overdue-invoice flow to the shipped column, then wrote one question beside the pilot proposal: “Which recurring decision would they trust us to own?”
That question gave the next conversation somewhere useful to begin.
Comments
No comments yet.