The design team can’t keep up with the rest of the company. Another hire or a different way of working?

The designer who fell behind

In February 2026, designer Neil Heiland wrote on LinkedIn that their team had changed its design process overnight. Heiland admits straight away that the change went against their own preferences: they liked designing slowly, caring about details, getting shadows just right. Reality turned out to be less romantic.

“engineering got faster, and I became too slow.”

Heiland connected an AI tool to the code repository and started designing directly with the components the product is built from. As they put it, the designer’s role shifted from making every screen to making the system those screens come from. That’s one person’s story, but teams we can look at more closely went the same way.

Sierra: nearly 20 times fewer designers than engineers

The AI in Design Report 2026 by Designer Fund and Foundation Capital takes a close look at several design teams. One of them is Sierra, a company building a platform of AI agents for customer service. Its design team is nearly 20 times smaller than the engineering organisation it designs for.

At first, Sierra’s designers did what instinct suggests. They used AI to polish production code themselves. It worked, but it didn’t scale. Designer Wayne Fan put it roughly like this: if a small team hand-finishes the work of a hundred engineers, it adds one more pair of hands and doesn’t raise the bar for anyone.

So the team moved its work upstream. It builds the components, patterns and structures engineers use to assemble a solid interface on their own, instead of fixing quality at the very end. It also changed how designs get handed over. Figma is now used mainly to show the overall concept, and designers go straight to working prototypes connected to the product environment. According to the report, the build, break and rebuild cycle that used to take days now fits into a single morning.

For the record: Sierra is still hiring designers and deliberately doesn’t automate everything. The report describes two designers spending hours drawing an icon by hand on a 16 by 16 pixel grid. So it’s hard to read this as a story about people replaced by AI. A small team stopped holding a big company back because it changed its own work.

Anthropic and designers who build

In the same report, Anthropic describes what Joel Lewenstein, its Head of Product Design, calls a timescale compression. AI writes most of the code there, so features that took weeks now ship in days. Design hasn’t sped up at the same rate. Exploration, decisions and polish increasingly set the pace of the product.

The response was to turn designers into builders. Every designer ships code in their first week, and designers are among the company’s heaviest AI users. They build their own tools for repetitive chores so more time is left for decisions.

For leadership, though, the most telling example in that case study is a different one. UX researcher Matt Gallivan used AI to build a weekly survey system and a product quality dashboard in about a week. He estimates a project like that might once have taken two or three researchers a couple of months.

What this changes about the hiring decision

The report also shows the other side of the shift. In its chapter on teams, 34% of respondents say collaboration between designers and the rest of the company has become messier, with roles and ownership less clear. The share who say there is less collaboration than before rose from 5% to 20% in a year. Access to tools doesn’t organise anything by itself. Someone has to design the process.

So before you sign another contract, check where your designers’ time goes today. If it’s mostly adaptations, post-release fixes, answering developers and manually policing consistency, a new hire will soon be stuck in the same queue. If you’re short of people for concepts and decisions because you’re launching a new product or entering a new market, a hire makes sense. Often you need both, just in that order: process first, then hiring based on data.

We sum up our AI design workflows service in one line: productivity grows, not headcount. We start with a process map and only then decide what AI should take over. For how the same shift looks from the developers’ side, read our post on design handoff in the age of AI.