Can you really run product discovery experiments more effectively with AI?
AI tools promise to accelerate product discovery — but do they? A critical look at where AI actually helps in discovery, and where the bottleneck remains human judgment.
“GenAI can enable cheaper, faster experimentation / discovery (it compresses the truth curve by reducing the cost of pretotyping style product experimentation techniques)” (Yours truly, in yesterday’s insight on how AI can really help you build better products)
This statement seems to have hit a nerve with reader Elad, who is product leader at a cybersecurity scaleup: “Not everyone can do this… New companies, sure. Larger, established companies are knee-deep in mountains of code, dependencies, and tech debt. Whenever you need to build an MVP that depends on your existing product, good luck…”
Elad is right. His comment echoes my initial assumption, which is that the jury is still out on how helpful “vibe coding” and GenAI code can be for large-scale brownfield/legacy software products in general.
The interesting opportunity I’m highlighting here is that GenAI, in general, and vibe coding, specifically, can make it easier to experiment in the pre-MVP stage, also referred to as Pretotyping.
Some of these experiments don’t even require coding. But a product team could run a landing page, a fake feature, an explainer video, and other prototyping techniques much faster using GenAI as an accelerator.
When you have built enough confidence in the truth curve that it makes sense to build an MVP, vibe coding should be considered as an approach to create a real, throwaway experiment intended to validate or invalidate.
Indeed, there will be some environments where the only way to conduct meaningful experiments will be to build something that integrates with your existing system or product; in such cases, vibe coding might not be beneficial.
In these cases, it may be beneficial to allocate more time to building higher conviction through pre-MVP discovery and experimentation techniques, before proceeding to expensive development in your real product code.
Yes, new product development in a greenfield is easier in many ways.
The advantage that product teams with existing products and customers have is precisely that - they have these customers. It should be easier for them to have conversations with these customers about problems, opportunities, needs, and jobs they’re trying to get done.
You win some, you lose some…
Have you used GenAI as an early pre-MVP experimentation technique? I’d love to hear about your experience…
Practical thinking on turning AI pilots, adoption, and portfolio work into business impact - by finding the constraint, changing the work, and proving value as you go.
Yuval Yeret helps product and tech leaders move from agile theater to evidence-informed delivery. Work with Yuval →
