The last uninstrumented function.
Design matters more than it ever has, and it has never been harder to run. Boundaries collapsed, output multiplied, and the people accountable for all of it are holding it together with memory, spreadsheets and goodwill.
I
Every mature function in a company runs on a system of record. Finance closes the books in an ERP. Sales lives in a CRM, to the point that a deal not in the CRM is treated as a deal that does not exist. Support runs on a ticket queue. Engineering has version control, which quietly became the most complete account of who did what that any function in the building possesses, and a category of software now exists purely to read it.
Each of those does the same job. It makes the work of the function legible to the people accountable for it, and to the business paying for it. Design never got one. It got extraordinary tools for making the work, which is a different thing entirely, and the gap went unnoticed for a long time.
II
What changed is scale and shape. In Figma’s 2026 report, drawn from 8,403 practitioners across ten markets, 57 percent said design has become more important as AI adoption grows. The share of developers doing design work rose from 44 to 60 percent in a single year, and the share of designers doing development nearly doubled. Forty-one percent said AI had meaningfully changed how their teams work together, against seven percent two years earlier.
The function did not simply grow. It dissolved into everything around it. There is more design happening, done by more people, in more places, than any design leader can observe directly. The org chart stopped describing the organization.
III
The same report found the thing worth sitting with: output is measurably up, and outcomes are not following. More gets made, and less can be shown for it. The alignment seams were always there. Abundance just made them load-bearing.
Meanwhile the questions a design leader answers have not changed and have not gotten easier. Where did the capacity actually go. Why is this slower than last quarter. What did this person contribute. What did any of it return to the business.
Those get answered by reconstruction: a spreadsheet before planning, a story before calibration, a week of asking around before a business review. Fifty-six percent of design system teams name staffing as their single biggest problem, and they are making that case without the numbers that would settle it.
A function becomes improvable at the moment it becomes legible. Design has been waiting a long time for that moment.
IV
The instinct is to reach for a dashboard. It will not work, and the reason is worth being precise about. A dashboard is a view onto data somebody already structured. Design’s data is not structured. It is a file, a thread, a critique, a study, a decision, scattered across six tools that each hold one piece and none of the connections between them.
The meaning lives in the connections. The study that prompted the file. The decision that came out of the critique. The person behind both. No tool in the stack can see past its own walls, so the connective work falls to a human, every time, from memory.
What is needed is not a report. It is a model of the organization itself: one that knows the work, the people and the links between them, learns how this particular company operates rather than how companies operate in general, and can act on that knowledge instead of only displaying it.
That model is what we are building. Agents read the work where it already happens, learn one organization’s teams and rituals and definitions in a model private to it, and take on the operational load that has always landed on design leaders instead of on design.
We do not think the next decade of design will be won by the organizations that produce the most. Production is solved, and getting cheaper every quarter.
It will be won by the ones that can see themselves clearly enough to get better on purpose.