AI is pushing product teams back to fundamentals – and exposing weak product management
For most of the past year, the conversation around AI in product teams has been fairly predictable. Faster prototyping. Faster delivery. Faster analysis. The underlying promise has been that if AI can remove friction from execution, then product teams will finally move at the pace they’ve always wanted to.
But that acceleration has a side effect that’s easy to miss. When engineers can ship faster, designers can prototype in hours, and AI can take on more of the operational work, execution stops being the main constraint. Something else takes its place.
According to Jarom Chung, VP of Product at Lucid, the new bottleneck is decision-making. As AI speeds up how products get built, teams are being forced to confront a harder question: do we actually know what we should be building, and why?
AI hasn’t just sped up product teams. It’s changed where they struggle. And in doing so, it’s quietly pushing teams back to fundamentals that many had drifted away from and exposing weak product management along the way.
When execution accelerates, weak product management becomes visible
AI tools are now everywhere in product workflows. Product teams can analyze feedback, usage data, and experiments almost instantly. At first glance, this looks like pure progress, but it also removes a kind of cover.
As Jarom explains, once delivery speed stops being the constraint, a more uncomfortable question arises, “What’s actually slowing us down now?”
When execution is no longer the bottleneck, the weaknesses in product thinking become much harder to hide. Roadmaps still get filled, backlogs still grow, and teams stay busy, but the decisions underneath all that activity, such as what to build, why now, and what risk is being taken, often feel surprisingly thin.
Jarom sees this clearly in how many product roles have drifted toward operational work.
“I think a lot of product managers get bogged down with the operational execution side, which is more like a product owner or project manager type of role,” he explains.
AI didn’t create this problem. It simply stripped away the buffer that used to hide it.
The fundamentals product teams can’t sidestep anymore
At its heart, product management was never meant to be about keeping tickets moving. It’s about making decisions when the answers aren’t obvious.
When Jarom talks about what really matters as AI reshapes product work, he keeps coming back to the basics. “It’s the strategy, it’s the product sense, it’s the innovation, it’s the validation, it’s the de-risking, it’s the solutioning,” he says. “Agents can help and will help with this over time. But right now, the need for strategic thinking is more human than AI.”
None of this is new. What is new is how little room there is to avoid it.
As AI takes on more of the mechanical work in summarising inputs, generating options, and accelerating analysis, the value of product managers shifts. Suddenly, they’re spending less time coordinating and more time thinking.
For many teams, this will be bigger than just a small adjustment. It will be a reset of what good product management actually looks like.
Why more data still doesn’t mean clearer decisions
There’s a comforting belief in product that if we just had better data, decisions would become more obvious. But in reality, more data usually means more trade-offs.
Jarom points to A/B testing as a simple example that most teams will recognize.
“I’ve probably seen hundreds and hundreds of tests… and I would say it’s probably less than 10 tests I’ve seen where all the metrics have all gone up.”
Something almost always dips. Engagement improves, but retention slips. Conversion goes up, while usage drops elsewhere. AI can surface these patterns faster and even suggest explanations, but it can’t necessarily tell you which trade-off matters most.
That’s where experience comes into play.
“In business, a lot of decisions are still subjective, and that’s where product sense and human intuition matter,” Jarom argues. “That intuition is built from seeing patterns over years. When I look at data and think, ‘something’s wrong here,’ it’s usually because I’ve seen this pattern before.
“At the same time, AI can help in places where decisions feel subjective but actually aren’t. You’ll see teams say, ‘this is a judgment call,’ when really one key metric has dropped in a way that will hurt the business. At that point, it’s no longer subjective – objectively, you shouldn’t do it.”
Roadmaps are thinking tools, not delivery checklists
Roadmaps sit right at the intersection of all this.
There’s a scientific side to roadmapping: analysing inputs, breaking work down, and understanding constraints. And then there’s the part that’s harder to codify, such as storytelling, empathy, and intent.
“It’s rare that all the data will align perfectly and tell you, ‘this is the right answer,’” Jarom says. “I think about data broadly – both qualitatively, what users are saying, feeling, and trying, and quantitatively, what the numbers show.
As AI increasingly handles the analytical heavy lifting, the human side of roadmapping becomes more important, not less.
The hidden risk of trying to do everything
One of Jarom’s most pointed observations comes with a warning label. “There’s this movement where PMs should be able to do everything. They should be able to code, they should be able to design, and they should be able to do product.”
In small teams, this can be a superpower, but in larger organizations it often creates a vacuum.
“If no one is thinking about those bigger decisions, more strategic things, and the innovation, then there’s still going to be that gap,” he says.
When product managers take on more execution work, the hardest parts of the role – typically discovery, prioritization, decision-making – are usually the first to get squeezed.
“I think designers should be doing more of the prototyping than product managers – that’s my hot take,” Jarom says. “If we just keep transferring more roles and responsibilities onto PMs, there’s still going to be a gap. Someone has to think about the bigger decisions like what the opportunities are, what the right solutions are, whether this is the right solution at the right time, how we iterate, and how we decide between conflicting data.”
AI makes this tension sharper, not softer. The more execution can be automated, the less sense it makes for product leaders to spend their time buried in it.
What will separate strong product teams next
Looking ahead, Jarom sees a clear divide forming between teams that thrive with AI and those that struggle.
It won’t come down to who adopts AI first. It will come down to what teams do with the time AI gives back. “Have AI automate as much as what you currently do in your role as possible and then use that time to be more innovative, more strategic, to validate and de-risk better,” Jarom argues.
Strong product teams will use AI to create space and then protect that space. Not for more output, but for better thinking.
Watch the webinar "Beyond the buzzwords: The 2025 product lessons you need to win in 2026" on demand.
AI isn’t replacing product management, it’s raising the bar
The important thing to remember is that AI isn’t making product managers obsolete. If anything, it’s removed many of the excuses for avoiding the hardest parts of the job.
As execution accelerates, fundamentals matter more. Strategy matters more. Judgment matters more.
The teams that succeed won’t necessarily be the fastest, but will be the ones willing to slow down, think rigorously, and return to the core of what product management was always meant to be.
Emma-Lily Pendleton
Read also
Experience the new way of doing product management
Experience the new way of doing product management