When AI compresses the time required for research synthesis, ideation, prototyping, and specification writing, it does not just change what individual designers do. It changes the structure of the teams that designers work within. The specialisations that made sense when each activity was expensive are being renegotiated. The handoffs that defined team boundaries are becoming less distinct. Understanding how this plays out -- and positioning yourself accordingly -- is one of the more strategically important things a designer can do right now.

How research has changed

The researcher-designer boundary has always been porous on high-functioning teams. AI has made it more so. When synthesis is fast, researchers can focus more of their time on the activities that genuinely require their expertise: participant recruitment and relationship management, session facilitation, the interpretation of data that requires contextual judgement. Designers can take on more of the synthesis work without it consuming the time it once did.

This is not a cost reduction argument -- it is not 'you need fewer researchers now'. The research that gets done should be getting better, because researchers are spending more of their time on the hard parts. The question for research teams is whether they are using AI to do more research or to do the same amount of research better. Both are valid; the answer should be a deliberate choice.

How the designer role has changed

The designer role is expanding upstream and downstream simultaneously. Upstream, designers are increasingly expected to engage with strategy and problem definition at a level that was previously the exclusive territory of product management. When the brief analysis and reframing work that used to take days can be done in hours with AI assistance, there is both the time and the expectation to do it as part of the design process rather than waiting for it to be handed over.

Downstream, the coded prototype territory I described in the Amazon Q piece is another example of the role expanding. Designers who can build working prototypes reduce the gap with engineering in both directions -- they understand implementation constraints better and engineers engage more genuinely with design intent when it arrives in a form that demonstrates understanding of the technical context.

The risk is spread too thin. An expanded role requires prioritisation. The most valuable thing a designer does is not everything they can now theoretically do -- it is the specific set of activities where their judgement, taste, and experience create the most leverage. AI creates the option to expand; it does not remove the obligation to focus.

How the PM-designer relationship is shifting

The traditional PM-designer division of labour -- PMs own the 'what', designers own the 'how' -- was always a simplification. AI is making it a less useful one. When designers can rapidly surface contradictions in requirements, generate multiple problem framings, and produce testable concepts in hours, the contribution of design to product strategy becomes more visible and more difficult to treat as a downstream activity.

The most effective PM-designer partnerships I see right now are ones where both parties are comfortable working in the messy upstream space together -- where design is contributing to problem definition and PMs are engaging with design thinking as a strategic tool rather than as a production capability. AI has made the case for that kind of partnership easier to make because the artefacts that demonstrate design's strategic value can be produced earlier and more frequently.

How to position yourself for the AI-era team

The AI-era designer is characterised less by tool mastery and more by thinking quality. The portfolio that demonstrates strong problem definition, clear commercial reasoning, and rigorous research practice will be more valuable than the portfolio that demonstrates Figma proficiency and visual polish. Both matter -- but the balance has shifted.

Building relationships across disciplines is also more important. When AI compresses the time-cost of individual activities, the bottlenecks in design processes increasingly show up in the coordination gaps between disciplines -- the moment where design is waiting for a product decision, or engineering is waiting for a spec that is comprehensive enough to work from. Designers who invest in reducing those coordination costs, through better communication and stronger cross-functional relationships, will be more valuable than those who invest only in their individual output.

The shift is not about becoming someone different. It is about doubling down on the parts of the job that were always the most valuable and are now, with AI removing much of the surrounding noise, the most clearly visible. Judgement, strategy, research rigour, communication. These were always what design at its best looked like. Now they are what design is expected to look like, consistently and demonstrably. That is a challenge worth rising to.