Linear's AI numbers make rollout an org chart problem, not an engineering one

Linear has published six years of its own workspace telemetry, covering tens of thousands of teams from before AI was widely adopted through to now ^[1]. Its headline finding is narrower and more useful than most vendor data: between January and June 2026, the share of users active on Linear's AI features more than doubled in every job function ^[2].

Product moved fastest, from 12% to 34% ^[3]. Go-to-market, the function Linear itself describes as furthest from the codebase, went from 5% to 18% ^[4]. That is a 3.6x increase in six months ^[2], and it puts go-to-market in June above where product sat in January ^[1]. A rollout plan that treats AI as a decision for the engineering org is already behind its own sales team.

Seniority does not soften the curve either. According to Linear, CEOs at companies of 201 or more people went from 9% to 36%, a 4x jump ^[5][3], the largest of any cut in the report, which Linear reads as senior leaders learning the tools by using them ^[5]. That cut leans on third-party company-size enrichment and covers fewer workspaces than the rest of the report ^[6]. Company size otherwise barely registers: adoption roughly tripled from startups to enterprises ^[7].

The composition of work has shifted with it. Two years ago fewer than one issue in a thousand was created by AI; teams now use AI to write just under half of everything created in Linear ^[8], an increase of more than 400-fold ^[4]. Meanwhile the human coordination around that output went up, not down. Time spent creating, triaging, and commenting rose in nearly every function, with engineering up roughly 17% on create and triage alone ^[9]. Founders swung hardest, up 17 minutes on creation and 26 on commenting, though Linear flags that cohort as small and noisy ^[10].

The most consequential line in the report is about what did not happen. Chatting with AI and delegating issues to agents are categories that did not exist a year ago and now appear in every function's week, with product leaning in hardest, and nothing else shrank to make room ^[11]. This is additive work so far, not substitution. Planning time also stayed flat: customer requests, docs, and projects held steady, which Linear takes as evidence that AI has changed how teams execute far more than how they decide what to build ^[12].

One structural change is visible at the output end. The share of product managers attaching pull requests rose from 3% to 10% over two years, and designers from 1% to 8% ^[13]. Linear only counts pull requests in repositories connected to Linear, so those are floors ^[14].

Two caveats do real work here. Linear cannot see AI usage outside Linear, so this is its customer base rather than the market ^[15], and functions are assigned by normalizing job titles, which carries error at the edges ^[16].

What to watch is whether the added layer stays added. If agents are genuinely absorbing execution, triage and commenting minutes should eventually fall rather than climb, and planning time should move. Linear frames this edition as a fixed point to measure the next one against ^[1], which is the comparison worth holding it to.