When the product shape is genuinely unknown, a roadmap is speculation dressed up as planning. You write down what you are going to build, you sequence it across quarters, and you commit a team to executing the sequence. The problem is that none of the inputs to that plan are evidence. They are guesses about what customers will want, projected against a market that does not exist yet.
A provocatype — a provocative prototype — is the alternative. It is a prototype built not to ship but to provoke conversation about how a tool ought to be used. Instead of building toward a predetermined spec, you create deliberately constrained or unusual interfaces — "what if the entire product was just a chat box?" or "what if there were no settings at all?" — and show them to customers to surface assumptions, preferences, and use cases that traditional requirements gathering misses.
The output of a provocatype is not a decision about whether to ship the prototype. It is a conversation. Customers react to the artifact in ways they cannot react to a slide deck or a Figma mockup. They tell you which constraints feel intolerable, which feel liberating, and which they did not know mattered until they saw the constraint imposed. The hidden preferences become legible.
This is especially powerful for agent companies, where the product surface is genuinely under-explored. Nobody has converged on the right way to expose long-running, autonomous, partially-autonomous, or supervised agent workflows. Roadmapping toward an answer assumes the answer exists. Provocatyping assumes you are going to discover the answer in the conversation.
I've argued elsewhere that the discipline of plucking a feather from each goose is the venue for this kind of work — a regular cadence with engaged partners, where the artifact you bring is not a roadmap update but a deliberately weird prototype. Roadmaps come later, after the primitive is clear. Until then, the right artifact is the one that provokes the conversation that tells you what to build.
Related Essays
Pluck a Feather From Each Goose
Following every engaged customer ends with three products instead of one. The discipline is to pluck a feather from each passing goose and follow none absolutely.
The Long Game in Agent Companies
Agent companies are marathons you sprint through. Frameworks commoditize, monetization pins to infrastructure, and proximity to real builders is the moat.
The Human Is No Longer the Integration Layer
The ERP is not going away. The CRM is not going away. The landing page is not going away. But the human as the operating system between them is. The agent takes that job now.
Key takeaways
- A provocatype is a prototype built to provoke conversation about how a tool ought to be used, not to ship features.
- Deliberately constrained or unusual interfaces — "what if the entire product was a chat box?" — surface assumptions that traditional requirements gathering misses.
- When the product shape is genuinely unknown, roadmaps are speculation. Provocatypes are evidence.
FAQ
How is a provocatype different from a normal prototype?
A normal prototype is built to validate that a planned feature works. A provocatype is built to provoke a reaction — usually by being deliberately constrained, unusual, or extreme. The output is not a decision about whether to ship; it is a conversation that surfaces hidden preferences and assumptions.
When should you use provocatypes versus a traditional roadmap?
Use provocatypes when the product shape is genuinely unknown — early category, no consensus on UX, customers who cannot articulate what they want. Use a roadmap once the primitive is clear and you are executing against it. Confusing the two is how teams waste 18 months.