Measure delivery from DORA to AI

LinearB connects your git, CI/CD, incident and AI tools, so you can see how work moves from first commit to production and whether AI speeds it up.

Cycle time, phase by phase

2d 8hAverage cycle time · sample data, last 30 days

Pickup and review take 24 of the 56 hours. Switch to AI-assisted PRs to see where that wait comes from.

Coding12hFair

Pickup9hFair

Review15hFair

Deploy20hGood

3.8 per dayDeploy frequency

6%Change failure rate

3h 10mMean time to restore

Trusted by the teams transforming software delivery

Find the bottleneck behind every DORA metric

LinearB calculates each DORA metric from your git, CI/CD and incident data. Choose a metric below and LinearB shows the leading indicators that move it and the work that needs attention first.

What moves deploy frequency

Smaller pull requests merge and deploy more often, so PR size and merge frequency are the first places to look.

Leading indicators

PR size182 linesFair

Merge frequency2.1 per dev a weekElite

Deploy time20hGood

Largest open pull requests

Refactor billing retries2,460linespayments-api · +1,840 −620

Move session store to Redis1,298linesplatform · +910 −388

Onboarding copy update84linesweb · +64 −20

See what AI changes on the delivery metrics you already trust

LinearB tracks adoption across 50+ AI tools, including GitHub Copilot, Cursor and Claude Code, and compares pull requests with and without AI on cycle time and DORA metrics.

All AI tools

Sample data · last 30 days

186AI active users

29%Suggestion acceptance

168AI-assisted pull requests

Delivery with and without AI · all tools

With AIWithout AI

Cycle timehours

64h

53h

Review timehours

26h

Change failure rate% of deploys

AI-assisted pull requests take 11 hours longer to merge than PRs without AI, and most of the gap is in review.

AI-generated pull requests are accepted at 32.7%, human-written at 84.4%.LinearB benchmarks · 8.1 million+ pull requests from 4,800 teams in 42 countriesRead the APEX framework guide

Start from leadership-ready dashboards

The LinearB dashboard catalog includes three dashboards built around one question each, with every widget explaining what it measures. Open one as it is, or clone it to adapt it for your teams.

Software Factory

LinearB Catalog

Efficiency81%Merged net of rework

Fully Agentic PRs14%of merged PRs

Review Load1.2×opened ÷ picked up

Merged PRs2,292PRs

Who Does the Work

Merged PRs by who did the work.

Only agentic 412Mixed 1,604Only human 276

Efficiency by AI Cohort

Efficiency of each cohort's PRs. Compare agent work with people's.

Human only88%

Mixed82%

Fully agentic71%

Which AI Tools Earn Autonomy

Each tool by efficiency (up is better) and AI cost per PR (left is cheaper). Size is PRs.

Give it more scopeImprove contextRestrict itTry cheaper models

Copilot

Cursor

Claude

Codex

Cost →↑ Autonomy earned

AI ROI

LinearB CatalogHow it works

Dev Annual Cost$140,000

AI Monthly Spend$10,000

Active Team Members40Devs

Cost Per Effective PR$1,153USD

Cost Per Effective PR by Month

Dev cost plus AI spend, divided by effective PRs merged.

Delivery Behind the Cost

You set two costs, and LinearB measures the rest from your delivery data.

1,310Total PRs Merged

2.5PRs Merged per Dev per Week

5.4%Rework Rate

1,240Total Effective PRs

Engineering Health

LinearB CatalogHow it works

Reviewer Imbalance4.2×Top 20% vs. the rest

Context Switch5.1Tickets / Dev / Day

Unplanned Work Share38%Unplanned items

Active on Weekends9%Devs

Weekly Review Load by Cohort

Weekly reviews per person for the busiest 20% of reviewers against everyone else.

Top 20% of reviewersBottom 80%

Development Drag

Where elapsed time piles up after the code is written.

In Review1d 16h

In Progress1d 4h

In QA22h

Deploy & Monitor9h

Sample data. All three dashboards are in the LinearB dashboard catalog.

See where you stand against 8.1 million+ pull requests

The data comes from 4,800 teams in 42 countries. LinearB shows which tier your organization sits in for each metric, so you know what good looks like before you set team goals.

Get the 2026 Software Engineering Benchmarks Report

Your organization by benchmark tier

Sample data

Needs focusFairGoodElite

Cycle time2d 8h

Pickup time9h

Review time15h

Deploy frequency3.8 per day

Merge frequency2.1 per dev a week

PR size182 lines

Change failure rate6%

Rework rate7%

Track the metrics your teams define

When a team measures something its own way, LinearB builds the metric from a plain-language description, shows the SQL behind it, and saves the chart to your dashboards.

See the Agentic Studio

Turn metrics into goals and automations

Every metric in LinearB connects to something you can do about it, from asking what's slowing it down to changing how pull requests get reviewed.

Ask what's slowing delivery down

Ask the LinearB AI assistant what's slowing your pull requests down, and get the teams, phases and pull requests behind the answer.

What's slowing down Payments' pull requests this month?

LinearBTwo reviewers handled 61% of Payments reviews, and PRs over 400 lines waited longest. Here are the five oldest.

Set goals for each team

Set goals for each team in LinearB, such as PR size or pickup time, and see on the goals dashboard where each team meets or misses them.

PR size under 200 linesPlatform

Pickup time under 12 hoursPayments

Automate the fix with gitStream

gitStream workflow automations assign code experts as reviewers, flag sensitive files and PRs without tests, and label every PR with an estimated review time.

Add retry budget to ledger sync

⏱ 5 min reviewCode expert assignedSensitive files

Match how your teams ship

Choose how LinearB detects releases, from tags or a dedicated release branch to the deployment API, override it for individual repositories, and get alerts in Slack or Microsoft Teams.

Release detectionTags · mobile-app uses a release branch

Alerts in SlackTeam channel when pickup time passes the goal

Customers

How engineering leaders are proving AI impact

Gartner® Magic Quadrant™

LinearB is a Leader

Matt CulverEngineering Chief of Staff, Super.comHow LinearB closes the AI delivery gap

Renan KosickiCTO, SkeeloRead the Skeelo story

Renan KosickiCTO, SkeeloRead the Skeelo story

Natively secure, obsessively compliant

Independently audited, with evidence of our controls available on request. LinearB participates in the EU-U.S. Data Privacy Framework.

GDPR: General Data Protection Regulation

Engineering metrics FAQs

All four. LinearB shows deploy frequency, change failure rate and mean time to restore (MTTR), and uses cycle time, from first commit to deployment, for lead time for changes. It also tracks the leading indicators behind them, such as PR size, pickup time, review time and merge frequency.

LinearB connects to your git provider, project management and CI/CD tools, and to 50+ AI tools. Incidents come from Jira, or from any incident tool through the LinearB Incident API.

LinearB tracks AI active users, suggestion acceptance and AI-assisted pull requests for each tool, then compares work with and without AI on cycle time and DORA metrics.

From LinearB's analysis of 8.1 million+ pull requests from 4,800 teams in 42 countries, published in the 2026 Software Engineering Benchmarks Report. LinearB shows where your organization sits against them in the app.

Yes. In the LinearB custom metrics builder (in beta), you describe a metric in plain language, check or edit the SQL it writes, and save the chart to a dashboard.

Yes. LinearB can detect releases from tags, from pull requests or merges to a dedicated branch, or from the deployment API, and you can override the method for individual repositories.