The canvas is dead

Figma is an undisputed market leader, a tool that has fundamentally improved the collaboration and hand-off process for digital product teams globally. I see its dominance not as an achievement, but as a strategic liability in waiting. The reality is that the core function of a canvas-based tool; drawing rectangles, placing components, and specifying micro-details, is a process that is about to become functionally obsolete.

Our industry is focused on efficiency, constraint, and measurable impact. Paying a senior designer a high salary to spend hours detailing a corner radius or ensuring component spacing is correct is a failed business model. It is a high-cost activity for a low-value, mechanical output. This realisation is the driving force behind the next necessary revolution in the design function: the death of the canvas.

The drag-and-drop ceiling

We must analyse the actual value generated by our design teams. The truly high-leverage work involves user behaviour analysis, system architecture, and strategic problem solving. The low-leverage work is the meticulous execution of an established design system within a static canvas.

We have hit a drag-and-drop ceiling. The current approach forces designers to operate as highly-paid pixel pushers. Even with advanced auto-layout features, the process remains one of manual rendering. Within the enterprise and live broadcast sector, where we deal with immense scale and high-stakes performance, this execution bottleneck is no longer tolerable. We require a system that generates a fully constrained, high-fidelity prototype from a simple, high-level brief. This is where Generative UI models are already beginning to prove their worth, capable of generating sophisticated component variations and full screen layouts with minimal input. The trend is accelerating, and we must organise our teams around this inevitability.

The 2025 reality: The shift to judgement

The market will soon be saturated with ‘Design-by-Prompt’ tools. For Enterprise UX teams, the day-to-day work will fundamentally change. The designer will no longer spend hours in a canvas detailing component states; they will spend minutes vetting and refining an AI-generated, high-fidelity prototype. The bottleneck moves instantly from production time to critique time.

I have always found that imposing a severe constraint, whether in product development or in my own creative pursuits, forces a higher level of strategic thinking. My experience with self-imposed constraints has taught me that true productivity comes from eliminating comfortable, low-impact tasks. Generative AI is imposing that necessary constraint on the design workflow now. The value is shifting from the ability to create a design to the ability to critique the generated output with strategic rigour.

The new design skill: Architecting the algorithm

If the machine handles the pixels, what is the high-value role of the designer? It is not mastering vector curves, but mastering design language models, or DLMs. The most valuable designers will be systems architects and prompt engineers.

Their core tasks will shift entirely:

Defining the system constraints

They will be responsible for defining the underlying component logic, architectural rules, and accessibility guidelines that the Generative UI model must adhere to. This is pure systems thinking.

Mastering the high-level prompt

Success will be defined by the quality of the high-level prompt provided to the model. The designer’s new craft is the precise communication of strategic intent. The goal is to generate an experience that meets all performance, P&L, and user behaviour requirements on the first pass, not to iterate endlessly in a canvas.

Our actionable advice to every designer must be direct: Stop treating tools like Figma as the final product! Your future job is not to design the button, but to design the algorithm that generates the button and the surrounding, constrained experience. The canvas is dead; our focus must be on the architecture that replaces it.