17 / 20Future of work

Designers and developers increasingly need to understand each other’s tools and decisions.

Designer over here.

Developer over there.

Handoff in the middle.

01Design and code are converging

That model becomes less convincing when design agents can work directly on a canvas, coding agents can read design systems, running interfaces can be captured back into Figma, and prototypes can become functional long before a formal handoff begins.

People who understand both design and development may be better placed to review work across the whole product. That is a professional judgment, not a measured pay premium for hybrid roles.

A designer who understands code can judge whether an interaction is structurally sensible before it reaches development.

A developer who understands design can recognise when technically correct output is experientially poor.

02Use AI across the workflow

That does not mean everyone should become equally expert in everything.

Breadth without depth is still a problem.

But the ability to cross boundaries becomes more valuable when AI handles more of the mechanical translation between disciplines.

03Direct the generated work

Figma’s current direction makes this explicit. MCP connects structured design context with coding agents. Code-to-canvas workflows bring production UI back into design. The design agent can act directly on the canvas.

In that environment, the key skill is increasingly direction.

Can you tell whether the generated component respects the system? Can you understand the layout and the DOM? Can you identify why the mobile experience fails? Can you adjust the design rather than merely accepting the first generated option?

This mirrors the kind of work that often happens in real projects.

04Solve product problems across disciplines

A responsive GitHub contribution calendar is not purely a design problem or purely a coding problem. Neither is an iframe that breaks at mobile width, a navigation system with awkward state, or a CMS component whose content model makes the interface harder to maintain.

The boundary is artificial.

AI is making that more obvious.

A more resilient career strategy may be to build judgment across the product, while keeping real depth in a craft.

05Keep depth and broad judgment

A resilient practitioner understands enough of the product to direct a model, evaluate its output and recognise where specialist depth is needed.

SourcesResearch trail

References used for this article

  1. 01Figma MCP
  2. 02Figma Code to Canvas
  3. 03Figma State of Design 2026