AI-assisted coding tools are enabling product managers to build interactive prototypes without touching backend systems, creating new friction on engineering teams. Eric Zakariasson, an engineer at Cursor focused on developer experience, said the shift demands that engineers "set clear expectations" with product teams.
Speaking at the AI Engineer Europe 2026 conference in remarks published Tuesday, Zakariasson argued that defining "what engineers kind of want from the product organization" is critical as roles blur. His comments arrive as AI code generators like Cursor lower the barrier to building functional demos.
"Maybe not vibe coding complete SaaS products is the most efficient thing," Zakariasson observed, referring to fully functioning apps built via AI prompts. He noted that product managers are spinning up prototypes rapidly, while engineers are left to make those demos production-ready.
The dynamic risks creating a bottleneck where non-engineers overestimate the readiness of AI-generated output. Engineers, Zakariasson suggested, must proactively communicate what is most helpful from product teams to avoid wasted effort.
Some industry observers counter that this tension is a natural productivity gain, not a crisis. They argue product managers building prototypes speeds iteration, and that clear workflows can be codified over time through shared tooling and practice.