2026-07 · 3 min read
The design of beginning
Impressive AI demos create attention. Good beginnings do something harder: they make a new capability feel close enough to try.
Most organizations assume that introducing a new capability is largely a question of showing what it can do.
I'm beginning to think it's a question of showing who it is for.
Capability demonstrations have become remarkably good. An agent completes a workflow. A prototype appears from a prompt. Hours of work disappear in seconds. The audience usually reacts the same way.
"Wow."
We tend to interpret that reaction as excitement. Sometimes it is. Sometimes it is distance.
A good demonstration shows people what is possible. It does not necessarily convince them that it is possible for them.
That difference explains something I've started noticing across many AI adoption efforts. The people who embrace a new capability first are rarely the ones who needed convincing. They were already curious. Already experimenting. Already looking for the next tool.
For everyone else, the demonstration often creates the opposite effect. The gap between where I am and what I just saw suddenly feels much larger. The technology doesn't feel closer. It feels further away.
The purpose of a first experience is not to demonstrate capability. It is to make capability feel reachable.
That changes what I think a beginning should look like. We often optimize for the most impressive example. The most autonomous agent. The most sophisticated workflow. The most surprising result. Yet those moments usually arrive far too early.
The first step is rarely about proving what the technology can do. It is about changing someone's belief that they can do it too.
The first barrier is psychological, not technical.
Perhaps this is also why adoption spreads in unexpectedly human ways. Organizations often invest heavily in presentations, documentation and showcase sessions. Those things matter. But I've repeatedly seen something much smaller create a greater shift.
A colleague leaning over and saying, "Try asking it this."
Not because the explanation is more complete. But because the distance suddenly becomes smaller. The technology no longer belongs to experts. It belongs to someone sitting next to you. The first step suddenly feels believable.
Thinking about it this way also changes the role of enablement. The goal is not simply to increase capability. It is to reduce psychological distance. To create enough early success that people stop asking, "Can I do this?" and start asking, "What else could I do?"
That transition is subtle. But it changes everything that follows.
Perhaps this is why so many AI initiatives struggle despite impressive technology. Organizations spend enormous effort demonstrating possibility. Far less effort designing beginnings.
No capability matters until someone believes it belongs to them. The most important design decision is often not how powerful the technology becomes. It is how possible the first step feels.