Closing the gap between design and experience

Jumping on the AI bandwagon, well, it is getting increasingly difficult to stay away from it :).

But I was determined not to put AI in the title of this article just to chase views. There are enough articles telling us how AI is changing design. I am more interested in something much more practical: what are we actually doing with it?

This question has been lingering in my mind for a while, particularly through conversations with my team about where AI fits into the way we work today. We have been experimenting with different tools and looking at where they can help us move faster. But somewhere in those conversations, a more interesting question started emerging.

Is AI actually helping us become better designers? Or are we simply using it to produce the same things we were producing before, only faster? More screens. More variations. More polished artifacts, produced in a fraction of a time. There is certainly value in that, but designers still need to create screens, communicate customer experiences, explore ideas, and bring clarity to what we are trying to build. I don't see that going away.

But I am beginning to think that productivity is the less interesting application of AI.

The bigger opportunity may be in using AI to close a gap that has existed in design for a very long time the gap between what we design and how the product actually behaves. And that is what I want to explore through this article.

In full disclosure, did I take AI's help in writing this article? Yes, I did. But not to sit back and let AI do its magic. I used it more like a colleague, someone to challenge, refine, and help shape an idea that was already mine into a better outcome.

Perhaps that, in itself, is part of the point.

It also made me think more about how we, as designers, are choosing to use AI today. There is no shortage of tools or possibilities, but are we using them to rethink how we design, or simply to accelerate the way we already work?

Most demos follow the same pattern — generate screens, create a flow, produce a prototype in a fraction of time. The excitement is understandable. But I also think we are focusing too much on one outcome: producing the same artifacts we were already producing before, just faster.

But has creating screens and flows really been our biggest challenge?

We already have tools that help designers do that faster and with greater accuracy. Design systems are a good example. Combined with tools like Figma, they have significantly improved the speed and consistency with which teams can move from an idea to a set of screens. I can vouch for that, having seen my own team operate and the difference it makes to their turnaround time.

So yes, AI can make us faster. But speed itself isn't new to design.

I think the more persistent challenge has been translation, the gap between what was intended, what was interpreted, and what users eventually experienced.

For years, artifacts have been at the center of the design process. Wireframes, flows, specifications, journey maps, decks, these became the language through which teams communicated intent. They help structure ambiguity and align cross-functional teams.

But somewhere along the way, we started mistaking the artifact for the outcome.

A screen, or even a set of screens, is not the experience.

A user flow represents behavior, but it is still a representation.

A prototype can demonstrate interactions, but it is still not the product.

All of these communicate intent. And wherever intent needs to be interpreted, there is room for something to get lost in translation.

One of the most consistent patterns I have observed across product teams is that misalignment rarely starts because people are careless.

• Design interprets a problem one way.

• Product reframes it through priorities.

• Engineering optimizes around constraints.

And eventually, users can experience something slightly different from what was originally intended.

Every handoff from design to product to engineering introduces some degree of drift.

The larger and more distributed the organization becomes, the more visible this problem gets. Teams spend enormous amounts of time refining flows, debating edge cases, reviewing specifications, and clarifying interactions, not because the ideas are weak, but because every artifact still requires some level of interpretation. And that is the part I think AI is beginning to make more interesting.

So perhaps the more interesting question isn't:

“How much faster can AI help us create?”

It is:

“How much closer can AI help us get to the actual experience?”

We already illustrate behavior through user flows. We already demonstrate interactions through prototypes. AI doesn't suddenly make those practices obsolete.

What it can potentially do is reduce the effort it takes to move from a screen, flow, or prototype to something that behaves much closer to what we actually intend to build.

And to me, that is a much more interesting form of acceleration.

It means designers can increasingly make tangible:

• How systems respond,

• How experiences adapt,

• How edge cases behave, and

• How recovery flows actually work.

  
  

The shift, then, isn't from designing screens to designing behavior. We have always thought about behavior. It is about shortening the distance between design intent and something people can actually experience.

That changes the conversation.

Instead of stopping at:

“What should this screen look like?”

or even:

“What should happen next?”

we can increasingly ask:

“Can we experience how this will actually work?”