Data reports · Edition 01 · June 2024 — August 2026
Tens of thousands of teams run product development inside Linear. Two years of that telemetry — from the first issue to the pull request that closes it — trace how AI actually changed software work. This is that story in nine charts, reinterpreted for everyone, not just engineers.
Linear sees the whole workflow behind building a product — the issue, the plan, the pull request — which makes this a rare window into the work itself rather than just token counts and code volume. The report answers three questions across that window: who is using AI, how it reshapes where teams spend their time, and whether it changes how much they ship.
One honest caveat, up front: everything here is measured inside Linear’s own paid customer base, not the market at large, and AI use outside the product is invisible to it. Read this as a fixed point to measure the next edition against — not a census.
Act I · Adoption
The first question is simple: who’s actually using this stuff? Between January and June 2026, the honest answer became — everyone, in every role, at every level, at every company size.
Chart 01 · Adoption
Between January and June 2026 the share of users active on Linear’s AI features more than doubled everywhere. Product climbed fastest, 12% to 34%. Even go-to-market — the function furthest from the codebase — went from 5% to 18%. Roles are classified by normalizing job titles, which adds noise at the edges, but the pattern is far too broad to be a labeling artifact.
Rendering chart… the data table below shows the same numbers.
| Segment | Jan 2026 | Jun 2026 | Change |
|---|---|---|---|
| Founder | 14% | 30% | +16pp |
| Engineering | 12% | 30% | +18pp |
| Product | 12% | 34% | +22pp |
| Design | 6% | 22% | +16pp |
| GTM | 5% | 18% | +13pp |
The read: no laggards. The slowest function (GTM +13pp) still nearly quadrupled. When the floor moves this fast, adoption is a property of the tooling, not the team.
Chart 02 · Adoption
Executives are personally active on AI at rates that match or beat their teams. CEOs at companies of 201+ people posted the largest jump of any cut in this report — 9% to 36% in six months — suggesting the most senior leaders are learning the technology by using it rather than reading about it. Company size comes from third-party enrichment, so this cut covers a smaller cohort than the rest of the report.
Rendering chart… the data table below shows the same numbers.
| Segment | Jan 2026 | Jun 2026 | Change |
|---|---|---|---|
| Founder, 201+ | 10% | 26% | +16pp |
| Founder, 51-200 | 15% | 27% | +12pp |
| Founder, 1-50 | 15% | 31% | +16pp |
| CEO, 201+ | 9% | 36% | +27pp |
| CEO, 51-200 | 15% | 25% | +11pp |
| CEO, 1-50 | 7% | 21% | +14pp |
| CPO, 201+ | 3% | 24% | +21pp |
| CPO, 51-200 | 10% | 26% | +15pp |
| CPO, 1-50 | 11% | 36% | +25pp |
| CTO, 201+ | 11% | 35% | +24pp |
| CTO, 51-200 | 12% | 28% | +16pp |
| CTO, 1-50 | 16% | 33% | +17pp |
The read: leadership isn’t delegating their AI curiosity. Every single executive row grew by double-digit points — CEOs of large companies 4×-ed their usage in half a year.
Chart 03 · Adoption
Normally, size is one of the best predictors of how fast an organization adopts new technology. Not here: adoption roughly tripled at enterprises and scrappy startups alike. Whatever is pulling people to AI, it isn’t filtered by org chart depth.
Rendering chart… the data table below shows the same numbers.
| Segment | Jan 2026 | Jun 2026 | Change |
|---|---|---|---|
| 1001+ FTE | 8% | 25% | +17pp |
| 201-1000 FTE | 9% | 27% | +19pp |
| 51-200 FTE | 9% | 25% | +16pp |
| 1-50 FTE | 8% | 23% | +14pp |
The read: the four rows are nearly on top of each other — 23–27% by June. AI adoption in 2026 behaves like infrastructure, not innovation.
Act II · Application
Adoption is the who. The how is more specific: teams pour more coordination into the system, AI starts writing half of what gets tracked, planning time stubbornly holds still — and a brand-new layer of work appears on top of everything.
Chart 04 · Application
Time spent creating, triaging, assigning, updating and commenting rose in nearly every function between June 2025 and June 2026 — engineering alone is up roughly 17% on creation and triage. Founders swing hardest (+17 min creating, +26 min commenting a month), though theirs is a smaller, noisier cohort. More work needs more coordination — and that coordination increasingly becomes the context agents act on.
Rendering chart… the data table below shows the same numbers.
| Segment | Jun 2025 | Jun 2026 | Change |
|---|---|---|---|
| Create & triage, Eng | 24 min | 28 min | +5m |
| Create & triage, Product | 38 min | 37 min | -1m |
| Create & triage, Design | 22 min | 25 min | +3m |
| Create & triage, GTM | 27 min | 31 min | +4m |
| Create & triage, Founder | 40 min | 57 min | +17m |
| Assign & update, Eng | 16 min | 19 min | +3m |
| Assign & update, Product | 26 min | 26 min | 0m |
| Assign & update, Design | 12 min | 15 min | +3m |
| Assign & update, GTM | 12 min | 15 min | +3m |
| Assign & update, Founder | 22 min | 29 min | +7m |
| Comment, Eng | 35 min | 40 min | +5m |
| Comment, Product | 48 min | 49 min | +1m |
| Comment, Design | 32 min | 34 min | +2m |
| Comment, GTM | 49 min | 55 min | +6m |
| Comment, Founder | 39 min | 64 min | +26m |
The read: only one bar in the whole chart sank (product create & triage, −1m). Everything else moved up — the system of record is being fed, not bypassed.
Exhibit · Two years, weekly
Two years ago, fewer than one issue in a thousand came from AI. This week in August 2026, agents and MCP clients authored 2,435 issues against 2,481 from people and integrations combined. At the current pace, AI will soon write the majority of everything tracked in Linear.
Rendering chart… the data table below shows the same numbers.
| Week of | Agents & MCP | People & integrations |
|---|---|---|
| 2024-06-03 | 0 | 605 |
| 2024-06-10 | 0 | 599 |
| 2024-06-17 | 0 | 582 |
| 2024-06-24 | 0 | 689 |
| 2024-07-01 | 0 | 602 |
| 2024-07-08 | 0 | 628 |
| 2024-07-15 | 1 | 621 |
| 2024-07-22 | 0 | 627 |
| 2024-07-29 | 1 | 650 |
| 2024-08-05 | 0 | 654 |
| 2024-08-12 | 1 | 624 |
| 2024-08-19 | 0 | 660 |
| 2024-08-26 | 1 | 650 |
| 2024-09-02 | 1 | 670 |
| 2024-09-09 | 1 | 690 |
| 2024-09-16 | 1 | 692 |
| 2024-09-23 | 1 | 725 |
| 2024-09-30 | 0 | 696 |
| 2024-10-07 | 1 | 726 |
| 2024-10-14 | 1 | 724 |
| 2024-10-21 | 1 | 741 |
| 2024-10-28 | 1 | 721 |
| 2024-11-04 | 1 | 760 |
| 2024-11-11 | 0 | 760 |
| 2024-11-18 | 1 | 795 |
| 2024-11-25 | 1 | 677 |
| 2024-12-02 | 1 | 765 |
| 2024-12-09 | 1 | 800 |
| 2024-12-16 | 1 | 770 |
| 2024-12-23 | 0 | 373 |
| 2024-12-30 | 0 | 460 |
| 2025-01-06 | 1 | 825 |
| 2025-01-13 | 1 | 878 |
| 2025-01-20 | 1 | 869 |
| 2025-01-27 | 1 | 920 |
| 2025-02-03 | 1 | 930 |
| 2025-02-10 | 1 | 924 |
| 2025-02-17 | 1 | 890 |
| 2025-02-24 | 1 | 934 |
| 2025-03-03 | 1 | 942 |
| 2025-03-10 | 1 | 974 |
| 2025-03-17 | 3 | 971 |
| 2025-03-24 | 3 | 984 |
| 2025-03-31 | 1 | 985 |
| 2025-04-07 | 1 | 999 |
| 2025-04-14 | 1 | 974 |
| 2025-04-21 | 1 | 994 |
| 2025-04-28 | 3 | 1,029 |
| 2025-05-05 | 5 | 1,037 |
| 2025-05-12 | 7 | 1,063 |
| 2025-05-19 | 9 | 1,042 |
| 2025-05-26 | 11 | 992 |
| 2025-06-02 | 18 | 1,095 |
| 2025-06-09 | 18 | 1,074 |
| 2025-06-16 | 28 | 1,064 |
| 2025-06-23 | 34 | 1,148 |
| 2025-06-30 | 35 | 1,092 |
| 2025-07-07 | 44 | 1,166 |
| 2025-07-14 | 40 | 1,137 |
| 2025-07-21 | 41 | 1,164 |
| 2025-07-28 | 45 | 1,177 |
| 2025-08-04 | 52 | 1,171 |
| 2025-08-11 | 50 | 1,206 |
| 2025-08-18 | 55 | 1,177 |
| 2025-08-25 | 47 | 1,225 |
| 2025-09-01 | 48 | 1,200 |
| 2025-09-08 | 46 | 1,299 |
| 2025-09-15 | 45 | 1,272 |
| 2025-09-22 | 45 | 1,294 |
| 2025-09-29 | 58 | 1,338 |
| 2025-10-06 | 62 | 1,352 |
| 2025-10-13 | 68 | 1,350 |
| 2025-10-20 | 65 | 1,382 |
| 2025-10-27 | 74 | 1,414 |
| 2025-11-03 | 85 | 1,461 |
| 2025-11-10 | 85 | 1,457 |
| 2025-11-17 | 91 | 1,436 |
| 2025-11-24 | 93 | 1,299 |
| 2025-12-01 | 122 | 1,473 |
| 2025-12-08 | 142 | 1,487 |
| 2025-12-15 | 147 | 1,526 |
| 2025-12-22 | 110 | 795 |
| 2025-12-29 | 139 | 808 |
| 2026-01-05 | 206 | 1,601 |
| 2026-01-12 | 273 | 1,725 |
| 2026-01-19 | 291 | 1,721 |
| 2026-01-26 | 323 | 1,806 |
| 2026-02-02 | 401 | 1,897 |
| 2026-02-09 | 451 | 1,901 |
| 2026-02-16 | 516 | 1,875 |
| 2026-02-23 | 599 | 2,048 |
| 2026-03-02 | 707 | 2,123 |
| 2026-03-09 | 794 | 2,170 |
| 2026-03-16 | 837 | 2,106 |
| 2026-03-23 | 916 | 2,297 |
| 2026-03-30 | 935 | 2,104 |
| 2026-04-06 | 1,038 | 2,063 |
| 2026-04-13 | 1,128 | 2,297 |
| 2026-04-20 | 1,209 | 2,173 |
| 2026-04-27 | 1,275 | 2,185 |
| 2026-05-04 | 1,382 | 2,238 |
| 2026-05-11 | 1,506 | 2,278 |
| 2026-05-18 | 1,597 | 2,271 |
| 2026-05-25 | 1,472 | 2,132 |
| 2026-06-01 | 1,542 | 2,270 |
| 2026-06-08 | 1,766 | 2,371 |
| 2026-06-15 | 1,652 | 2,256 |
| 2026-06-22 | 1,728 | 2,372 |
| 2026-06-29 | 1,799 | 2,265 |
| 2026-07-06 | 2,078 | 2,532 |
| 2026-07-13 | 2,143 | 2,465 |
| 2026-07-20 | 2,195 | 2,396 |
| 2026-07-27 | 2,348 | 2,357 |
| 2026-08-03 | 2,435 | 2,481 |
The read: watch the violet line cross the dashed amber one. Around July 2026 they meet — from ~0 to half of all intake in 26 months.
Chart 05 · Application
In a year when nearly everything else in this report moved, time on customer requests, docs and projects held steady. Planning practice varies wildly between teams and much of it happens in conversation before it lands anywhere, so the average blends heavy planners with light ones. What the flatness suggests: AI has so far changed how teams execute far more than how they decide what to build.
Rendering chart… the data table below shows the same numbers.
| Segment | Jun 2025 | Jun 2026 | Change |
|---|---|---|---|
| Customer requests, Eng | 1 min | 1 min | 0m |
| Customer requests, Product | 3 min | 4 min | 0m |
| Customer requests, Design | 1 min | 1 min | 0m |
| Customer requests, GTM | 4 min | 4 min | +1m |
| Customer requests, Founder | 2 min | 3 min | +1m |
| Docs & projects, Eng | 3 min | 3 min | +1m |
| Docs & projects, Product | 13 min | 14 min | +1m |
| Docs & projects, Design | 4 min | 5 min | +1m |
| Docs & projects, GTM | 3 min | 3 min | +1m |
| Docs & projects, Founder | 7 min | 8 min | 0m |
The read: every change rounds to 0–1 minutes. Whatever AI is doing to software work, it hasn’t (yet) touched how long teams spend deciding what the work even is.
Chart 06 · Application
Chatting with AI and delegating issues to agents are categories of work that didn’t measurably exist a year ago. Both now show up in every function’s month, with product leaning in hardest (5 min chatting). And critically — nothing else shrank to make room. AI landed on top of the job, not instead of it.
Rendering chart… the data table below shows the same numbers.
| Segment | Jun 2025 | Jun 2026 | Change |
|---|---|---|---|
| Agent issues, Eng | 0 min | 1 min | +1m |
| Agent issues, Product | 0 min | 1 min | +1m |
| Agent issues, Design | 0 min | 0 min | 0m |
| Agent issues, GTM | 0 min | 0 min | 0m |
| Agent issues, Founder | 0 min | 2 min | +2m |
| Chat with AI, Eng | 0 min | 2 min | +2m |
| Chat with AI, Product | 0 min | 5 min | +5m |
| Chat with AI, Design | 0 min | 3 min | +3m |
| Chat with AI, GTM | 0 min | 3 min | +3m |
| Chat with AI, Founder | 0 min | 4 min | +4m |
The read: zero to 1–5 minutes per person per month, everywhere, in twelve months. Small numbers — but they measure a category that was exactly 0 before, growing on top of an unchanged workload.
Act III · Output
The last question: does any of this change what ships? Three views — who’s attaching pull requests, how many pull requests exist at all, and which teams are doing the accelerating.
Chart 07 · Output
The share of product managers attaching pull requests rose from 3% to 10% in two years; designers went 1% to 8%. Only PRs in repositories connected to Linear count, so these numbers are floors, not ceilings. The people who used to describe a change increasingly ship it themselves.
Rendering chart… the data table below shows the same numbers.
| Segment | Jun 2024 | Jun 2025 | Jun 2026 | Change (2 yrs) |
|---|---|---|---|---|
| Founder | 11% | 12% | 23% | +12pp |
| Engineering | 20% | 22% | 34% | +14pp |
| Product | 3% | 3% | 10% | +7pp |
| Design | 1% | 2% | 8% | +7pp |
| GTM | 1% | 1% | 3% | +2pp |
The read: every function’s line rises — but note the slopes steepen after 2025. Designers 8×-ed their shipping share as coding agents went mainstream.
Chart 08 · Output
Pull requests opened per paid workspace held roughly level for the first twelve months — then bent upward through 2026 as model quality and adoption climbed together. The count is PRs opened, not merged, and an opened PR says nothing about the change’s value. But the inflection is hard to miss.
Rendering chart… the data table below shows the same numbers.
| Week of | Change vs Jun 2024 |
|---|---|
| 2024-06-02 | 0% |
| 2024-06-09 | +9% |
| 2024-06-16 | +10% |
| 2024-06-23 | +3% |
| 2024-06-30 | +8% |
| 2024-07-07 | -4% |
| 2024-07-14 | +8% |
| 2024-07-21 | +7% |
| 2024-07-28 | +8% |
| 2024-08-04 | +7% |
| 2024-08-11 | +6% |
| 2024-08-18 | +3% |
| 2024-08-25 | +10% |
| 2024-09-01 | +10% |
| 2024-09-08 | +5% |
| 2024-09-15 | +12% |
| 2024-09-22 | +10% |
| 2024-09-29 | +14% |
| 2024-10-06 | +8% |
| 2024-10-13 | +11% |
| 2024-10-20 | +9% |
| 2024-10-27 | +18% |
| 2024-11-03 | +8% |
| 2024-11-10 | +15% |
| 2024-11-17 | +11% |
| 2024-11-24 | +17% |
| 2024-12-01 | 0% |
| 2024-12-08 | +16% |
| 2024-12-15 | +17% |
| 2024-12-22 | +10% |
| 2024-12-29 | -58% |
| 2025-01-05 | -50% |
| 2025-01-12 | +6% |
| 2025-01-19 | +15% |
| 2025-01-26 | +13% |
| 2025-02-02 | +15% |
| 2025-02-09 | +19% |
| 2025-02-16 | +21% |
| 2025-02-23 | +17% |
| 2025-03-02 | +21% |
| 2025-03-09 | +19% |
| 2025-03-16 | +26% |
| 2025-03-23 | +26% |
| 2025-03-30 | +23% |
| 2025-04-06 | +17% |
| 2025-04-13 | +24% |
| 2025-04-20 | +11% |
| 2025-04-27 | +10% |
| 2025-05-04 | +7% |
| 2025-05-11 | +14% |
| 2025-05-18 | +21% |
| 2025-05-25 | +22% |
| 2025-06-01 | +9% |
| 2025-06-08 | +22% |
| 2025-06-15 | +16% |
| 2025-06-22 | +12% |
| 2025-06-29 | +22% |
| 2025-07-06 | +8% |
| 2025-07-13 | +16% |
| 2025-07-20 | +16% |
| 2025-07-27 | +16% |
| 2025-08-03 | +13% |
| 2025-08-10 | +9% |
| 2025-08-17 | +5% |
| 2025-08-24 | +10% |
| 2025-08-31 | +8% |
| 2025-09-07 | +4% |
| 2025-09-14 | +11% |
| 2025-09-21 | +10% |
| 2025-09-28 | +8% |
| 2025-10-05 | +9% |
| 2025-10-12 | +9% |
| 2025-10-19 | +9% |
| 2025-10-26 | +9% |
| 2025-11-02 | +14% |
| 2025-11-09 | +15% |
| 2025-11-16 | +13% |
| 2025-11-23 | +16% |
| 2025-11-30 | +1% |
| 2025-12-07 | +17% |
| 2025-12-14 | +17% |
| 2025-12-21 | +14% |
| 2025-12-28 | -48% |
| 2026-01-04 | -54% |
| 2026-01-11 | +10% |
| 2026-01-18 | +22% |
| 2026-01-25 | +22% |
| 2026-02-01 | +27% |
| 2026-02-08 | +32% |
| 2026-02-15 | +36% |
| 2026-02-22 | +33% |
| 2026-03-01 | +50% |
| 2026-03-08 | +49% |
| 2026-03-15 | +54% |
| 2026-03-22 | +55% |
| 2026-03-29 | +58% |
| 2026-04-05 | +41% |
| 2026-04-12 | +46% |
| 2026-04-19 | +60% |
| 2026-04-26 | +66% |
| 2026-05-03 | +67% |
| 2026-05-10 | +80% |
| 2026-05-17 | +91% |
| 2026-05-24 | +95% |
| 2026-05-31 | +85% |
| 2026-06-07 | +106% |
| 2026-06-14 | +113% |
| 2026-06-21 | +111% |
The read: a year of noise around the baseline, then a staircase: +50% by March 2026, +111% by June. Something structurally changed in early 2026.
Chart 09 · Output
A fixed cohort of teams with a coding agent connected went from 21 to 65 pull requests a week — nearly a tripling. Teams without one went from 8 to 10. The levels aren’t directly comparable — agent-linked teams were higher-output before agents existed — but each cohort against its own baseline tells a clean story: nearly all the growth sits on the agent side.
Rendering chart… the data table below shows the same numbers.
| Week of | Coding-agent teams | Traditional teams |
|---|---|---|
| 2024-06-02 | 21 | 8 |
| 2024-06-09 | 24 | 8 |
| 2024-06-16 | 24 | 9 |
| 2024-06-23 | 22 | 8 |
| 2024-06-30 | 24 | 9 |
| 2024-07-07 | 21 | 8 |
| 2024-07-14 | 24 | 8 |
| 2024-07-21 | 24 | 8 |
| 2024-07-28 | 24 | 9 |
| 2024-08-04 | 24 | 8 |
| 2024-08-11 | 24 | 8 |
| 2024-08-18 | 23 | 8 |
| 2024-08-25 | 25 | 8 |
| 2024-09-01 | 24 | 8 |
| 2024-09-08 | 24 | 8 |
| 2024-09-15 | 25 | 9 |
| 2024-09-22 | 25 | 8 |
| 2024-09-29 | 26 | 9 |
| 2024-10-06 | 25 | 9 |
| 2024-10-13 | 26 | 8 |
| 2024-10-20 | 25 | 8 |
| 2024-10-27 | 26 | 9 |
| 2024-11-03 | 25 | 8 |
| 2024-11-10 | 27 | 9 |
| 2024-11-17 | 26 | 8 |
| 2024-11-24 | 28 | 9 |
| 2024-12-01 | 23 | 8 |
| 2024-12-08 | 27 | 9 |
| 2024-12-15 | 28 | 9 |
| 2024-12-22 | 26 | 8 |
| 2024-12-29 | 10 | 3 |
| 2025-01-05 | 11 | 4 |
| 2025-01-12 | 25 | 8 |
| 2025-01-19 | 27 | 9 |
| 2025-01-26 | 27 | 8 |
| 2025-02-02 | 28 | 8 |
| 2025-02-09 | 29 | 9 |
| 2025-02-16 | 30 | 9 |
| 2025-02-23 | 29 | 9 |
| 2025-03-02 | 30 | 9 |
| 2025-03-09 | 30 | 9 |
| 2025-03-16 | 30 | 9 |
| 2025-03-23 | 31 | 9 |
| 2025-03-30 | 31 | 9 |
| 2025-04-06 | 30 | 8 |
| 2025-04-13 | 32 | 9 |
| 2025-04-20 | 28 | 8 |
| 2025-04-27 | 28 | 8 |
| 2025-05-04 | 28 | 8 |
| 2025-05-11 | 29 | 8 |
| 2025-05-18 | 32 | 9 |
| 2025-05-25 | 31 | 9 |
| 2025-06-01 | 28 | 8 |
| 2025-06-08 | 31 | 8 |
| 2025-06-15 | 31 | 8 |
| 2025-06-22 | 30 | 8 |
| 2025-06-29 | 32 | 8 |
| 2025-07-06 | 29 | 8 |
| 2025-07-13 | 32 | 8 |
| 2025-07-20 | 31 | 8 |
| 2025-07-27 | 32 | 8 |
| 2025-08-03 | 32 | 8 |
| 2025-08-10 | 32 | 8 |
| 2025-08-17 | 31 | 7 |
| 2025-08-24 | 33 | 8 |
| 2025-08-31 | 32 | 8 |
| 2025-09-07 | 31 | 8 |
| 2025-09-14 | 34 | 8 |
| 2025-09-21 | 34 | 8 |
| 2025-09-28 | 34 | 8 |
| 2025-10-05 | 35 | 8 |
| 2025-10-12 | 34 | 8 |
| 2025-10-19 | 34 | 8 |
| 2025-10-26 | 34 | 8 |
| 2025-11-02 | 36 | 8 |
| 2025-11-09 | 36 | 8 |
| 2025-11-16 | 35 | 8 |
| 2025-11-23 | 37 | 8 |
| 2025-11-30 | 31 | 7 |
| 2025-12-07 | 37 | 8 |
| 2025-12-14 | 38 | 8 |
| 2025-12-21 | 37 | 8 |
| 2025-12-28 | 16 | 3 |
| 2026-01-04 | 13 | 3 |
| 2026-01-11 | 35 | 7 |
| 2026-01-18 | 40 | 8 |
| 2026-01-25 | 39 | 8 |
| 2026-02-01 | 42 | 8 |
| 2026-02-08 | 44 | 8 |
| 2026-02-15 | 46 | 9 |
| 2026-02-22 | 44 | 8 |
| 2026-03-01 | 49 | 9 |
| 2026-03-08 | 50 | 9 |
| 2026-03-15 | 51 | 9 |
| 2026-03-22 | 50 | 9 |
| 2026-03-29 | 52 | 9 |
| 2026-04-05 | 48 | 9 |
| 2026-04-12 | 49 | 9 |
| 2026-04-19 | 54 | 9 |
| 2026-04-26 | 55 | 9 |
| 2026-05-03 | 55 | 8 |
| 2026-05-10 | 57 | 9 |
| 2026-05-17 | 60 | 10 |
| 2026-05-24 | 62 | 10 |
| 2026-05-31 | 57 | 9 |
| 2026-06-07 | 65 | 10 |
| 2026-06-14 | 63 | 9 |
| 2026-06-21 | 65 | 10 |
The read: same two years: agent teams +210%, traditional teams +25%. The weekly gap between the two cohorts is now 55 PRs — and still widening.
A closing note
The clearest signal in the whole dataset: AI adoption and shipping acceleration move together. Whether that output converts to business outcomes, no dashboard can yet say — but the coincidence of the two curves is difficult to argue with.
Adoption is blurring roles: senior leaders doing hands-on IC work and adopting AI aggressively to do it, non-engineers committing code. Directionally, the suggestion that everyone in an organization is becoming a builder now has numbers behind it.
The gains didn’t show up as time saved. Time on existing tasks held while AI work appeared as a new layer, so total effort is going up, not down. Teams are working more, not less — efficiency became capacity, not leisure.
Pull requests measure motion, not worth. It beats measuring tokens — a mechanical refactor burns mountains of them while a one-line fix burns none — but the next edition’s job is the full lifecycle: from token spend to outcomes.
The source report uses aggregated product telemetry from Linear: AI conversations, agent sessions, issue activity, comments and pull requests, covering paid workspaces and their users only. All metrics are reported in aggregate — broad patterns, never individual behavior.
This is an independent visual reinterpretation of How Teams Build: AI usage patterns in software teams (Edition 01), by Tim Qi, Head of Data at Linear (2026). Every number was extracted verbatim from the public data tables published at linear.app/data on August 20, 2026; the charts, copy, and design are new. This page is not affiliated with or endorsed by Linear.