Over the last year, AI-enabled design tools have meaningfully changed how quickly teams can produce artifacts. We can generate flows, screens, variants, and prototypes at a pace that would have been unthinkable eighteen months ago. That progress is real.

I’ve been thinking about what comes after the speed. What I’m seeing in executive conversations is genuine excitement about velocity. What we don’t have yet is shared language for what kind of speed we’re actually buying. Most AI enthusiasm today is enthusiasm for production acceleration: how fast teams can make things. The harder question is whether that production speed is translating into decision speed, meaning faster and better decisions grounded in user reality.

Production speed reduces the cost of making artifacts. Decision speed reduces uncertainty at the moment choices get made. One improves throughput. The other improves outcomes. Confusing them creates a subtle but expensive risk: teams feel confident sooner without actually being more correct.

There’s an assumption embedded in a lot of AI conversations. Faster making leads to faster iterating, which leads to better outcomes. I wonder if we’re missing the actor in that loop. When we say “iterate faster,” who are we iterating with? Users, or ourselves?

AI dramatically increases internal velocity: how quickly teams align, revise, and move ideas forward inside the building. What it does not automatically increase is external learning, meaning how quickly our assumptions get validated with real customers, in real contexts, with real consequences.

That gap matters. Internal velocity without external learning doesn’t create insight. It creates rehearsal. Teams move quickly and confidently…in directions that may or may not map to reality.

The gap shows up wherever an organization depends on functions that carry signal from outside the building: UX, customer success, sales, support, research. These roles exist to create grip — the connection to reality that keeps speed from becoming drift. I’ve spent twenty-six years on the UX side of this, and the pattern is consistent. When production timelines compress, the functions that slow down to listen get pressure to speed up and match. Here’s where I’d pause and reflect, because their value was never in keeping pace. It’s in keeping contact.

So the real question is what we’re multiplying. Insight, or opinion?

When AI accelerates decisions grounded in validated user understanding, that pays off. When it accelerates decisions that haven’t been tested outside the room, I could see that becoming confidence debt. And confidence debt, like technical debt, always comes due, usually at a less convenient time. I’ve watched this play out across fintech, insurance, and enterprise software. Teams ship in half the time, celebrate the velocity, then spend three quarters untangling assumptions that were never tested. The rework is quiet (which is what makes it easy to miss). A feature that doesn’t get adopted. A flow that requires constant support intervention. A roadmap that keeps revisiting the same problem because it was never actually solved.

Here’s what I would say to leaders looking for a simple diagnostic: if an iteration loop doesn’t include a real user signal, we’re rehearsing.

For any AI-accelerated design work, we should be able to answer three questions. What assumption are we testing? Who outside this room can confirm or refute it? How quickly will that signal come back?

Those are governance questions more than UX questions. When AI compresses production timelines, the assumptions embedded in what we’re building get locked in faster. That makes the decision about when and how we validate a strategic choice rather than a research preference. Who owns that decision, and how it gets resourced, sits with leadership.

AI has made us fast at moving inside the organization. I wonder how we might get equally fast at moving toward reality. Paired with learning, the speed pays off. Without it, we just get confidently wrong sooner.

Curious what others are seeing here. What’s your experience?

I’ve spent twenty-five years leading UX across fintech, cybersecurity, enterprise software, and insurance tech. And I keep seeing the same pattern: good UX work—solid research, thoughtful design, real effort—fails to shape decisions in meaningful ways.

It’s not because the work is bad. It’s because most organizations aren’t ready for what the work asks of them.

I’ve watched teams genuinely invest in discovery, engage in design exploration, even agree with what they’re seeing—only to quietly move forward with the original plan anyway. Not because they don’t care. Not because they don’t “get UX.” But because uncertainty is uncomfortable, and most organizations are built to resolve discomfort quickly.

UX introduces pause. Organizations reward momentum.

When UX work creates tension—between speed and rigor, roadmap and reality—what I usually see isn’t outright rejection. It’s erosion. Insight gets acknowledged but softened. Design intent gets diluted. Decisions get reframed as “pragmatic” when they’re really just familiar.

Early in my career, I thought the answer was more research. Clearer artifacts. More “actionable” deliverables. I’ve learned that volume isn’t the issue. Capacity is.

Capacity to sit with ambiguity. Capacity to question assumptions without panicking. Capacity to let understanding actually change direction.

You can see the breakdowns when that capacity isn’t there:

• Research lives in decks, not in decisions
• Design intent makes sense in concept but falls apart in delivery
• Teams want solutions before they’ve aligned on the problem

Those aren’t process failures. They’re human ones.

Organizations, like people, develop coping mechanisms under pressure. Metrics, velocity, and quick decisions start to feel safer than slowing down to think. Certainty becomes more comfortable than clarity.

What I’ve learned works differently

UX doesn’t need to fight harder to be heard. UX needs to function as infrastructure, not output.

I think about it as three connected capabilities:

Understanding is where we listen—really listen—to what’s true for users, even when it’s inconvenient. Research surfaces patterns, tensions, and signals we might prefer not to see. This breaks down when teams hear insights but aren’t ready to sit with them.

Interpretation is where understanding becomes shared meaning. Design frames the problem, makes tradeoffs visible, and turns insight into intent. This breaks down when we rush to solutions before unresolved questions are actually resolved.

Follow-through is the hardest part. It’s about protecting intent when pressure arrives—deadlines, scope, competing priorities. This is where leadership matters most. When stress rises, meaning either holds or dissolves.

UX rarely breaks at handoffs. It breaks at sense-making gaps—when research is acknowledged but not integrated, when design is appreciated but not protected, when delivery optimizes speed over understanding.

The real work

The organizations I’ve seen do this well share one thing: leadership willing to tolerate discomfort long enough for insight to become understanding.

UX maturity isn’t about process sophistication or headcount. It’s an organization’s ability to make meaning together—without panic, ego, or false certainty.

That’s not a tooling problem. It’s a leadership one.

And after twenty-five years, I’m more convinced than ever that building that capacity is the actual work.