The Quiet Anxiety in Every Product Team

Bjorn did not shy away from naming the elephant in the room. Product managers are watching AI compress timelines, automate workflows, and blur the lines between disciplines. The fear is real: if everyone can build, does the PM role even survive?

“We are the bottleneck in this whole story and that's also partly why you are here tonight. You see everything moving so quickly and you're thinking I can't catch up yet.”
- Bjorn Vuylsteker

That admission landed hard. But what followed was not a doom narrative. Instead, Bjorn drew a clear line between two mindsets: coding-to-code, where someone takes pride in the act of building itself, and product thinking, where the goal is creating value that other people experience. The PM who falls into the first camp will always feel threatened by faster tools. The PM who lives in the second camp will see AI as the most powerful lever they have ever had.

“I am a product person. That's what I realized a few years ago. I love to make value for other people and for other people to experience that value.”
- Bjorn Vuylsteker

What Stays: The Irreplaceable PM Skills

The first half of Bjorn's framework focused on what endures. These are the capabilities that AI cannot replicate and that become even more valuable as execution speeds up. If you are a PM investing your growth energy anywhere, these are the places to double down.

De-risking delivery

Faster building does not mean lower risk. In fact, the speed itself introduces new failure modes. The ability to identify risks early, design for graceful failure, and sequence work to reduce uncertainty remains a core PM discipline. AI can ship features in hours, but it cannot tell you which features carry hidden dependencies or which assumptions have not been validated.

Maximizing business value

This one sounds obvious, but Bjorn made a sharp point: when build cost drops toward zero, the opportunity cost of building the wrong thing becomes the dominant concern. The PM who can consistently identify the highest-value work and say no to everything else is more critical than ever. That judgment cannot be delegated to a model.

True customer insight

“Really understanding the customer, digging more into the customer because guess what human-to-human communication is not really something that an AI has. It is something that is human.”
- Bjorn Vuylsteker

This was the most emphatic point in the section. AI can analyze survey data and cluster feedback, but it cannot sit across from a frustrated user and read the tension in their shoulders. It cannot notice that a customer says one thing but does another. Genuine empathy and observational skill remain uniquely human, and PMs who sharpen them will pull ahead.

lightbulbObservation as a Skill

Observation is a skill, not a passive activity. Bjorn stressed the importance of capturing what you see and structuring it so that AI tools can access it later. The insight is human; the leverage comes from making that insight available at scale.

What Shifts: The PM Skills Being Rewritten

The second half of the framework is where things get uncomfortable. These are not new skills appearing from nowhere. They are existing PM capabilities that are transforming so significantly that the 2024 version and the 2026 version barely resemble each other.

Orchestration looks completely different

PMs have always been orchestrators. But the shape of that orchestration is changing fast. Instead of coordinating human teams through stand-ups and status updates, you are increasingly directing AI agents that work with a growing set of tools. At the same time, you are still orchestrating customers and teammates, just through different touch points and at a different pace.

“Now you're orchestrating an AI that's working with all these tools. You don't have to go around yourself.”
- Bjorn Vuylsteker

Living and breathing software

In a world where prototyping happens in real time and user feedback can be incorporated within the same session, the relationship between PM and product becomes much more intimate. You are no longer reviewing a spec, waiting for a build, and testing a release. You are inside the loop, watching the product take shape and making micro-decisions continuously.

This raises a new challenge that Bjorn called out directly: knowing when to stop. Where do you declare a checkpoint? Where do you say 'this is done for now' before continuing? The old sprint boundary handled that question by default. In an AI-accelerated workflow, the PM has to create those boundaries deliberately.

Data literacy and guardrails

The final shift is perhaps the most consequential. As products become more intelligent and more autonomous, the PM needs to understand where data flows, what the product learns from that data, and where to set hard boundaries. This is not a nice-to-have analytics skill. It is a core responsibility that touches privacy, ethics, and product integrity.

warningGuardrails Are Not Optional

If you do not understand your product's data flows, you cannot set meaningful guardrails. And in 2026, a product without guardrails is a liability waiting to surface.

The New PM Toolkit: Putting It Together

Taken together, Bjorn's framework paints a picture of a PM role that is simultaneously more strategic and more hands-on. You spend less time managing process and more time in the product itself. You spend less time writing specs and more time talking to customers and structuring what you learn. You spend less time coordinating humans and more time directing AI agents toward the right goals.

Strengthen your customer insight practice. Schedule more user conversations, not fewer. Capture observations in structured formats that AI can leverage.
Learn to orchestrate AI agents. Understand the tools, their capabilities, and their limitations. Your value is in knowing what to ask for, not in doing the work yourself.
Develop checkpoint discipline. In fast-moving AI workflows, you are the one who decides when something is ready and when it needs more iteration.
Build data literacy as a core skill. Know where your product's data comes from, where it goes, and what boundaries need to exist.
Practice saying 'this is done.' Speed creates a temptation to keep iterating forever. The PM who can declare done and move to the next high-value problem wins.

Two PMs, Same Product, Different Approaches

Imagine two PMs running the same AI-powered product. PM A treats AI as a faster version of their old workflow: write spec, hand off, review output. PM B embeds themselves in the AI loop, provides real-time context from customer conversations, sets clear checkpoints, and knows exactly what data the product is using. PM B ships better products in less time because they have adapted their toolkit to match the new reality.

The PM Role Is Not Shrinking. It Is Deepening.

The anxiety that Bjorn named at the start of his talk is understandable. When the landscape shifts this fast, uncertainty is the rational response. But the framework he offered is a clear antidote: the core of what makes a PM valuable has not changed. Understanding customers, de-risking delivery, and maximizing business value are as important as they have ever been.

What has changed is the surrounding toolkit and the pace at which it operates. PMs who adapt their orchestration skills, develop fluency with AI agents, and build real data literacy will find that the role is not being diminished. It is being elevated. The upgraded PM does not compete with AI. They wield it with the kind of judgment and taste that only comes from deeply understanding the people they are building for.

infoAbout This Talk

This article is based on the concluding section of Bjorn Vuylsteker's talk at The Product Consortium's Vibe Coding Workshop, held on March 12, 2026. The full event covered hands-on AI prototyping, knowledge management for product teams, and the evolving PM skill set.

The Upgraded PM: What Stays and What Shifts in 2026

Download the Slides

The Quiet Anxiety in Every Product Team

What Stays: The Irreplaceable PM Skills

De-risking delivery

Maximizing business value

True customer insight

lightbulbObservation as a Skill

What Shifts: The PM Skills Being Rewritten

Orchestration looks completely different

Living and breathing software

Data literacy and guardrails

warningGuardrails Are Not Optional

The New PM Toolkit: Putting It Together

Two PMs, Same Product, Different Approaches

The PM Role Is Not Shrinking. It Is Deepening.

infoAbout This Talk

This article is for Consortium members

Key Takeaways

Related Editorials

The Vibe Coding Trap

The Commoditization of Execution: Why 'Taste' is the New PM Moat

Context is the New Code: Building an AI Knowledge Base

Zero to Validated Prototype in 10 Minutes

Stop Prompting, Start Orchestrating: The Era of Agentic Workflows