Uncovering AI Needs When Users Don't Know What They Need

As I move deeper into AI strategy work, I'm reminded of something deceptively simple: people cannot ask for what they do not yet understand. And in the context of AI, that gap between possibility and perception is vast.

This is not a failure. It is the path.

Tools don't land, stories do

Recently, we introduced a capability designed to enhance model performance through better quality data. The value was there. The potential was real. But we led with the how instead of the why. The result? Misalignment. The users saw a tool. We saw a transformation.

The lesson: people engage with stories before they engage with solutions. Before you show someone what AI can do for them, you need to help them imagine a world in which they are different: faster, more confident, less encumbered by the work that bores them.

The discovery problem

Traditional discovery methods assume users can articulate their needs. User interviews, surveys, workshops: all of these presuppose a level of AI literacy that most organisations simply don't have yet.

The better approach is observational and speculative. Watch how people work. Notice where they slow down, where they copy and paste, where they make decisions that feel like they shouldn't need a human. Then bring AI possibilities to those moments, not as abstract capabilities, but as concrete "what if this was handled for you?" scenarios.

Designing for the unknown

The best AI product work we do at Kablamo starts not with requirements but with a question: what would have to be true for this team to work differently?

That question opens space for imagination before it demands specification. It gives people permission to want things they don't yet know how to ask for.

From there, the work becomes collaborative. We're not extracting requirements; we're building a shared understanding of what's possible, together.

Originally published on the Kablamo blog.