When another AI video platform ships the feature at the centre of your next quarter, do not rewrite the roadmap on Friday. Use the weekend to test whether customers valued the feature itself, the problem it solved, or the advantage you expected it to create.
In 1985, Intel’s leaders faced a harder version of that decision. Japanese manufacturers were beating the company in memory chips, the business on which Intel had built its identity. Andy Grove asked chairman and CEO Gordon Moore what new leaders would do if the board replaced them. Moore said they would get out of memory. Grove’s response was simple: why not walk back through the door and do it themselves?
When the old roadmap describes the old company
Grove recounts the exchange in Only the Paranoid Survive. Intel’s eventual move toward microprocessors now reads like an obvious strategic turn. At the time, it meant loosening the company’s grip on the product category that had defined it.
That is the useful part of the story. Grove and Moore were not choosing between a good roadmap and a bad one. They were deciding whether the assumptions beneath the roadmap still held.
An AI video founder can face that question between Friday afternoon and Monday morning. The competitor’s release may match the planned headline feature, produce better output, cost less to run, or reach users through a distribution channel the founder cannot copy. The team still has tickets assigned, designs approved and a launch date in the calendar. None of those facts prove the work should continue.
A roadmap carries momentum. Customer evidence deserves more weight.
Separate the customer problem from the planned feature
My first move would be to write down the exact promise the feature was supposed to fulfil.
“Ship automatic scene continuity” is a product commitment.
“Help a small marketing team produce a usable campaign without correcting every shot” is a customer problem.
Those statements lead to different Monday decisions. If the new platform has solved scene continuity for everyone, rebuilding the same capability may consume runway without creating a reason to choose your product. If customers still struggle to turn generated clips into an approved campaign, the underlying problem remains open. The roadmap may need a different feature, workflow or market boundary.
This distinction matters most for small teams. A pre-seed founder in Accra, Lagos, Cape Town, Berlin or San Francisco cannot treat three months of engineering as a symbolic response to a competitor. The cost appears in delayed customer work, fewer experiments and less cash available for the next decision.
The same principle shaped my view in Should We Rebuild Monday’s Campaign When a Better AI Video Model Arrives?. A model release can change what is technically possible overnight. It does not automatically reveal what customers will pay you to own.
Make Monday’s decision reversible
By Monday, I would want a one-page decision record with three columns.
The first would list what became true on Friday. Only observable facts belong there: the competing platform shipped the capability, users can access it, and the output covers some or all of the planned use case.
The second would list what remains uncertain. Can customers rely on the output in their existing workflow? Does the platform handle the markets, devices, approval steps or distribution constraints that matter to them? Will buyers pay for a product built around the remaining gap?
The third would define the smallest test that can change the decision. That could mean showing the competitor’s release to five current prospects, replacing the planned build with a manual service for one customer, or testing whether the team’s proprietary workflow still improves the final result.
Then I would choose among three bounded moves: continue for one short milestone, redirect the team toward the unresolved customer problem, or stop the work. Each choice needs a review date and evidence threshold. “Keep going while we think” quietly turns uncertainty into another quarter of spending.
This is also why hiring ahead of demand can become dangerous. A roadmap feels harder to question once salaries depend on it, as I explored in Should I Run Friday Payroll or Hire an AI Engineer Before Testing Demand?.
Decide what the company still owns
Intel’s 1985 question stripped away history, pride and internal momentum: what would a new leadership team do with the facts available now?
An AI founder can use the same device without copying Intel’s answer. Imagine the board replaced you on Monday. The incoming founder has the current bank balance, customer conversations, product usage and Friday’s competitor release. Would that person fund the same quarter?
If yes, continue, but state the remaining advantage precisely.
If no, change the roadmap before the team spends another week protecting it.
The practical task for Monday morning is small. Open the roadmap, mark every item whose rationale depended on being first, and require fresh customer evidence before that item returns to active work.
Comments
No comments yet.