Save
What I didn’t do
I have not vibe-coded, crafted the perfect prompt, created a skill to use, or [insert any trending thing right now].
Not because I’m not interested — believe me, it’s tempting. But my focus has been elsewhere. It simply hasn’t aligned with my goals and needs.
Not yet.
For the last 6 months, I’ve been exploring how to build better experiences between humans and technology. Through understanding both the technology and human behavior.
So instead of building, I’ve read academic research. I’ve taken notes. I’ve developed my own interpretations. I’ve been developing principles to guide product teams on how to build for appropriate trust, adoption, chatbots, and prompts.
While personally fulfilling, it’s been rewarding to see how this helps product teams. Because these experiences happen before all the layers get added on.
Did it really change what I’m doing?
Yes and no.
Yes, in terms of tools and how I use them. No, because I am still deciding the direction, interpreting insights, and translating them into something doable.
At first, it’s intimidating to reevaluate your tasks, your skills, and how you add value when AI gets introduced into your workflow. But as I found more clarity on my value independent of tools, I moved into this new phase with a little more confidence.
It really comes down to a check-in with yourself
Why was I in tech in the first place? Why am I working in product?
For me, it has always been about giving people the best possible experience. To do that now, I need to understand this new relationship more deeply.
The “invisible” is quite visible
Let’s start with the easier one. How we feel can be expressed. Verbally, through what we say. Visually, through facial expressions and gestures. Have you ever watched someone tap a button that doesn’t work? It’s normally more than once. We can take those emotional cues, adapt our screens, and fix that button.
But what about something harder, like building appropriate trust?
That’s a big one.
I mean, how would we even measure that? People can say they trust it, or they don’t, but is that enough? My hunch: no, it’s not.
One way to approach it is to understand what drives trust in the first place. Why do some people over-rely while others under-rely?
It turns out there are pre-existing conditions and biases that influence trust. And understanding those is what helps us go from an invisible concept to something visible and actionable in a product.
Here’s an example. A user might have an automation bias, blindly trusting recommendations either because they lack the expertise to evaluate accuracy or because past positive experiences have made them complacent. One way to look at this is through cognitive forcing. A confirmation step that highlights high-stakes decisions and prompts the user to review before proceeding.
Depending on how users perceive the AI feature, we either build in more trust mechanisms or become more transparent about its limitations. The invisible becomes baked into the product. The screens you see and interact with and the way they function — that’s all designed intentionally. Not copy-pasted from somewhere else.
The goal is to actively create a better experience, aligned with both business goals and the people using it.
For me, this has formed into something specific. Translating theory from academic research, which, I’ll confess, is usually a hard read, into actionable frameworks and principles that guide product teams within their own context.
We are making the invisible visible.
Context, context, context
We can build the same experience for everyone. But that doesn’t mean it will work for your customers. Same humans. Different needs. Different environments. Different behavior.
We could’ve just had stairs between floors, but we don’t.
Some of us have mobility needs.
Some of us are claustrophobic.
Some of us are active.
Some of us have strollers.
And most of us, if we’re honest, are just a little lazy.
The way we get there might look different, but we all reach the next floor.
(And while we’re here, why is baby clothing on the top floor when most of your customers probably have a stroller and active wear on the ground floor? Asking for a friend.)
The same logic applies to your product. The challenge is finding the balance. Not so generic that no one feels it was built for them, and not so complex that no one can get started.
Yes, please build, but build intentionally
Building and creating have become easier. But building contextually and intentionally? That’s where the real challenge is.
We are responsible for how people perceive AI in our products. We build it. We design it. We feed it.
If a user isn’t trusting or using your latest AI feature, it’s not the AI at fault. It’s how we built and designed it to work.
The article originally appeared on Medium.
Featured image courtesy: Immo Wegmann.
- The piece states that building AI features is easy. But building them on purpose, turning invisible human behaviors like trust and bias into deliberate design choices, is where the work lives.
