AI product builder vs PM and Designer: who does what in ai?

Senior manager data & ai transformation at Converteo, Erik Perrier is an experienced leader in data and ai transformation. He assists organizations in defining their product strategy and managing teams dedicated to innovation in the field of ai and agentics.

Key takeaways

• the product manager defines the “why”, the designer defends the user, the ai product builder materializes the vision into a testable artifact—three complementary, not competing, roles.
• the limit of the pm and the designer when facing ai is the same: they can conceive the intention, but struggle to confront their hypotheses with the real behavior of a probabilistic system.
• the ai product builder precisely bridges this gap: they translate a vision into a functional mvp to de-risk the investment and accelerate learning.

Artificial intelligence is not only transforming technology—it is breaking down the traditional boundaries of roles within the product team. Faced with systems that learn and act autonomously, the historical responsibilities of the product manager, product owner, and product designer are being profoundly challenged.

A new profile is emerging, not to replace them, but to fill a critical void between strategic vision and technical reality. This profile is the ai product builder. Understanding who they are—and especially what they are not—is essential for building ai products that work and last.

The product manager: strategist and guardian of value

The role of the product manager—or the ai product owner—remains the strategic pillar of the team. They are the guardian of the “why”. Their mission is to define the product vision, ensure its alignment with the company’s business objectives, understand the market, and build the roadmap. They are responsible for the final value delivered to the user and the company.

With ai, this role is more important than ever. The pm must assess whether a problem truly requires an ai solution or if a simpler approach is sufficient. However, they quickly hit a wall: the fundamental uncertainty of the technology. They can define a vision, but they have difficulty concretely evaluating its technical feasibility and its probabilistic behavior—which can hinder their ability to make quick and informed decisions.

The product designer: advocate for the user experience

The ai product designer is the user’s advocate. They focus on the experience, the journey, the ergonomics, and the interaction. Their mission is to make the product usable, understandable, and enjoyable. To do this, they use tools like interviews, empathy maps, and prototyping software to create wireframes and user journeys.

In an ai context, their role becomes more complex. How do you design a conversation? How do you prototype a system whose responses are non-linear and sometimes unpredictable? Classic design tools, designed for deterministic journeys, show their limits. The designer can conceive the intention of an interaction, but they struggle to materialize and test the agent’s real behavior.

The ai product builder: the architect-builder who bridges the gap

The ai product builder emerges to precisely fill the gap between the pm’s vision and the experience imagined by the designer. Their distinctive skill is to immediately translate a vision into a tangible and testable artifact. They are a hybrid profile that merges product vision, technical mastery, and a strong “build” capability.

AI product builder vs. product manager

Where the pm defines the need and prioritizes the roadmap, the ai product builder builds the first functional ai mvp to validate or invalidate hypotheses. They don’t just “think” the product—they “make” it on a small scale to de-risk the investment and radically accelerate learning. They are obsessed with the speed of materialization.

AI product builder vs. product designer

Where the designer creates static wireframes or visual prototypes, the ai product builder designs functional and interactive prototypes with no-code ai tools. They are not only interested in the appearance, but in the agent’s real behavior. They use “vibe design” to capture user reactions to a living system—something no static prototype can reproduce.