AI is changing Product far beyond making us faster. As building becomes easier and teams work closer together, judgement, learning and knowing what is worth building become the real advantage. And as software starts understanding our intentions, the role of Product itself may be about to change.
I have spent a large part of my career somewhere between product and software development.
And for most of that time, building a digital product followed a fairly predictable path. We understood the problem, defined requirements, designed the experience, handed it over to development, and eventually put it in front of users.
Of course, real product development was never quite that linear. There were discussions, changes, compromises and plenty of going backwards. But the boundaries between the disciplines were clear.
AI is making those boundaries much less clear.
At first, the change looked mostly like speed. Writing a document became faster. Research could be summarised in seconds. Developers could generate code and designers could produce alternatives almost instantly.
That was impressive, but I don’t think speed is the most interesting thing happening anymore.
What I find much more interesting is that the distance between Product, Design and Engineering is getting smaller.
The distance is changing because of AI
I don’t believe these roles are merging into one, and I don’t think they should.
Specialisation still matters. A product manager does not suddenly understand software architecture because they can generate code. A developer does not become a product designer because they can generate an interface. And being able to produce something is very different from having the experience to know whether it is good.
But we can now go much further into each other’s worlds.
Instead of Product spending days defining something, passing it to Design and eventually passing that to Engineering, we can put something rough in front of each other much earlier and ask: Does this actually make sense?
That sounds like a small operational improvement. I think it is much bigger than that.
A technical constraint appears after an experience has already been designed. An edge case emerges halfway through development. Something that made perfect sense in a requirements document feels completely wrong once you can actually click through it.
The goal, then, isn’t to eliminate teams. Ownership still matters. The goal is to make those teams work closer.
When everything becomes easier to build
There is another side to this that I think product leaders need to pay attention to.
Since AI makes building dramatically easier, then our biggest limitation starts moving somewhere else.
We can generate more ideas. More screens. More prototypes. More code. More features.
But more has never been the goal of Product.
The difficult question has always been: What is worth building?
There is also a strange side effect to how good these tools are becoming: they can make progress look more real than it is.
A prototype can look convincing before the problem behind it has been properly understood. A specification can sound complete while important questions remain unanswered. We can generate an impressive amount of research summaries, ideas, tickets and documentation and still not be any closer to knowing what the right decision is.
In fact, more output can sometimes create more noise.
This is something I think product teams need to become particularly conscious of. Polish is not evidence and output is not progress.
AI can analyse customer feedback and find patterns that may have been missed. But it wasn’t in the room with that customer. It doesn’t know that one comment changed the entire tone of the conversation.
It can produce a very convincing answer to a question we should perhaps never have been asking in the first place.
When execution becomes easy, judgement becomes more critical.
The fundamentals suddenly matter more
There is an irony in all of this.
The more sophisticated our tools become, the more I find myself coming back to the very basic product questions.
Who are we building this for? What problem are we actually solving? How are we going to solve it?
None of these questions are new.
Research is not new. Empathy is not new. Experimentation, product judgment, understanding trade-offs and talking to users are certainly not new.
AI hasn’t made any of them obsolete. It has simply made it possible to move from a question to something tangible much faster. And that creates both an opportunity and a risk. Because speed without judgment simply allows us to make mistakes faster.
A team can now spend a morning building something that previously took a week. That’s valuable if we use the afternoon to put it in front of someone, learn that our assumption was wrong and change direction.
For me, a significant opportunity behind AI is learning sooner, not producing more.
There will still be moments when we need to slow down. There will still be technical decisions that belong with engineers and experienced decisions that require design expertise. AI doesn’t remove accountability.
But if I look further ahead, I think something more fundamental may happen.
The product team of the future may need fewer layers between an idea and the people capable of making it real.
Teams may become smaller, more multidisciplinary and more autonomous. Not because one person suddenly does five jobs, but because fewer steps are required for five different people to understand and contribute to the same problem.
That changes what I would look for in a great product person, designer or engineer too.
Deep expertise will still matter. But so will curiosity beyond your own discipline, the ability to understand context, to challenge an answer, and to recognise when something that looks finished isn’t actually solved.
The tools we use will keep changing and the boundaries between our roles probably will too but core principles remain.
What happens when that intention becomes the starting point of the product?
For years, we have designed the paths people take through software. We have entered a period where software increasingly finds that path for them.
The first signs are already here. Instead of opening several websites, comparing options, filling in forms and moving through predefined flows, we can now tell an AI what we are trying to achieve. It can find the information, compare the options and, with our permission, begin taking actions on our behalf.
This is where I believe Product may evolve in a much more interesting direction. If AI increasingly handles the steps between intention and outcome, our job is no longer only to design the perfect sequence of screens. We also have to design what the system should understand, what decisions it can make, when it needs permission, how it earns trust and where the human must remain in control.
And if that direction continues, the future of Product may be less about designing every path a user can take, and more about designing the rules, trust and boundaries.
*Stavri Vasileiou is Head of Product at Silicon Dali. Her path into Product began with a background in Mechanical Engineering and evolved through software development and UX. Today, she leads Product at Silicon Dali, shaping complex, business-critical systems across Forex, Crypto and iGaming
