AI is moving beyond the code editor

Linear’s latest data report uses aggregated product activity visible inside Linear to show how AI is being adopted across software teams, not only by engineers writing code but also by product, design, executives, and go-to-market roles.

The company can observe work that happens inside Linear, including AI conversations, issue delegation to agents, issue activity, comments, and pull requests. It cannot see AI use in external tools, so the report should be read as a view into Linear’s customer base rather than a full market survey.

Adoption is broad, including leadership

Between January and June 2026, the share of users active on Linear AI features more than doubled in every function. Product roles rose from 12% to 34%, the fastest increase in the report. Even go-to-market teams, which are typically farther from the codebase, moved from 5% to 18%.

Executives are also using AI directly. Among CEOs at companies with 201 or more employees, AI activity increased from 9% to 36% over six months, the largest jump among the cuts Linear highlighted. The company says role classification is based on normalized job titles, which can introduce some edge-case errors, while company-size data comes from third-party enrichment and covers fewer workspaces.

Key signals include:

• AI-active share more than doubled across every function;
• adoption roughly tripled across company sizes;
• company size appears to matter less than usual for this technology cycle.

More AI has not meant less work

Linear’s data suggests AI is adding a new layer of activity rather than replacing existing collaboration. From June 2025 to June 2026, time spent creating, triaging, and commenting increased in nearly every function. Engineering time on creation and triage alone rose by roughly 17%. Founders showed larger swings, adding 17 minutes on creation and 26 minutes on commenting, though Linear notes that this cohort is smaller and noisier.

Planning activity was comparatively stable. Time spent on customer requests, documents, and projects did not move much, even as many execution-related measures rose. Linear interprets this as evidence that AI has so far changed how teams execute more than how they decide what to build.

New work categories have appeared: chatting with AI and delegating issues to agents. These activities did not exist in the same way a year earlier, and they now show up across functions, with product teams leaning in most. Importantly, other activity did not shrink to make room.

Output is rising, especially with coding agents

AI is now responsible for a large share of issue creation inside Linear. Two years ago, fewer than one in a thousand issues was created by AI. Today, AI authors just under half of all new issues, and Linear says it may soon create more than people and integrations combined.

Pull request activity is also changing. Product managers attaching pull requests rose from 3% to 10% over two years, while designers rose from 1% to 8%. Since Linear only counts repositories connected to its system, these figures are presented as floors rather than ceilings.

Across paid workspaces, pull requests opened per workspace are up 111% from a June 2024 baseline. Linear counts opened PRs, not merged PRs, and an opened PR does not prove that the change was valuable. Still, the increase is visible.

Coding agents appear to explain most of the acceleration. In a fixed cohort of paid workspaces, teams that connected a coding agent increased weekly pull requests from 21 to 65 over two years. Teams without one moved from 8 to 10. Linear cautions that agent-connected teams were already higher-output before coding agents existed, so absolute levels are not directly comparable; the stronger comparison is each group against its own baseline.

The next question is value, not volume

The report is useful because it moves beyond token counts. A token is a basic unit of text processed by an AI model, but token volume is a weak proxy for value: a mechanical refactor can consume many tokens, while a meaningful bug fix may not.

Linear’s data points to a real shift in software organizations. Leaders are doing more hands-on individual-contributor work, non-engineers are increasingly attaching code changes, and teams are feeding more structured context into systems that agents can act on. The idea that more people inside a company are becoming “builders” looks directionally credible.

But the evidence also shows that teams are working more, not clearly working less. AI usage, coordination, and output are all rising together. The industry’s next measurement challenge is to connect this activity to outcomes: better products, faster resolution of customer needs, higher code quality, or stronger business performance.