42 → 72
The average pull request grew from 42 to 72 lines of code in a year. Bigger changes are harder to review well.
Engineering intelligence for AI-assisted teams
Prove that AI pays off, without breaking production.
One engineering health score per team. AI spend, delivery and reliability on one page, in euros.
From €12 per collaborator a month · no seat minimum · no credit card
One engineering health score per team: DevEx, OpsEx and AgileEx, with what changed and what to do next, written every morning. Try it in the free tour →
The problem
Coding got faster. Shipping got riskier.
3×
Under high AI adoption, incidents per pull request more than tripled. Median time in review rose fivefold.
Faros AI, The Acceleration Whiplash, April 2026, 22,000 developers
79% vs 92%
Even at the best 10% of organisations, 79% of pull requests opened by an AI agent merge within 30 days. For people it is 92%.
The paradox
You are paying for speed you cannot see yet.
28×
The rise in median quarterly AI spend per organisation in one year, in the tech sector.
- $1.5K → $44KMedian quarterly AI spend per organisation, a year apart. DX, Q2 2026
- +23%More pull requests once AI agents arrived, across 718 companies. Chen & Stratton, Harvard, 2026, Jellyfish data
- No changeIn epics resolved. More code went in; no more finished work came out. Same study
What it answers
Three questions every engineering leader now gets asked.
01
Is AI paying off?
- Adoption per tool: Claude Code, GitHub Copilot, Cursor, Codex and Gemini.
- Cost and return in euros per AI-assisted merge request, from licences and metered usage.
- Delivery speed of AI-assisted work against the rest, compared inside each repository.
AI Impact
Adoption per tool, cost and return in euros per AI-assisted merge request, delivery speed and change failure rate against the rest. Try it in the free tour →
02
Is AI breaking things?
- Quality guardrails: change failure rate and time to restore for AI-assisted work.
- Error budgets per service, so reliability slips show up before customers notice.
- DORA metrics over time, set against the share of AI-assisted work.
Error budgets per service: how much unreliability is left this period, and which incident used it up. Try it in the free tour →
03
Is it wearing people out?
- Who reviews whom, and which repositories rest on one person.
- Repositories nobody owns any more, and work that has been in progress for weeks.
- Kudos: the recognition colleagues give each other, counted per team.
Who works on which repository and who reviews whom: repositories that rest on one person, and repositories nobody owns. Try it in the free tour →
One score
One number per team, explained every morning.
The engineering health score is the average of three thirds. Each third is the average of its areas. Every morning Agile Analytics writes what changed and what to do next: in the app, in Slack and by e-mail.
DevEx
How developers work
AI adoption · Repository health · Recognition
OpsEx
How production holds up
Delivery · Reliability
AgileEx
How plans become shipped work
Flow · Sprint predictability · Hours efficiency
Overview
One engineering health score per team: DevEx, OpsEx and AgileEx, with what changed and what to do next, written every morning. Try it in the free tour →
Ask
Ask your engineering data a question.
Type it in plain language, in the app or by mentioning the bot in Slack. The answer comes from your own data and links to the page it came from.
Need the numbers somewhere else? An API key reaches every number through the REST API today.
Ask
Ask about your teams in plain language, in the app or by mentioning the bot in Slack. Answers come from your own data. Try it in the free tour →
Why Agile Analytics
Built for teams, independent, and European.
Team-first
Scores are per team. Every insight shows how it was written, so a lead can check it before acting on it.
Independent
Reads GitHub, GitLab, Bitbucket, Jira, Slack, Google Cloud Monitoring and CloudWatch. Not owned by a platform vendor, so it has no tool to favour.
European
Euro pricing from €12 per collaborator a month. No seat minimum, a 60-day trial, a GDPR data processing agreement, data processed in the EU. A Dutch company.
Setup
Four steps. Your first report arrives the next morning.
Connect your Git host
GitHub, GitLab or Bitbucket. Merge requests, reviews and deployments come from here.
Connect Jira
Sprints, work in progress and flow come from your boards.
Add monitoring and AI tools
Google Cloud Monitoring or CloudWatch for error budgets. The AI tools you pay for, for cost and adoption.
Add Slack and invite your teams
The morning summary goes to Slack and by e-mail. The bot answers questions there too.
Work and code
- Jira
- GitLab
- GitHub
- Bitbucket
- REST API
Where reports go
- Slack
- E-mail digest
Monitoring
- Google Cloud Monitoring
- CloudWatch
AI tools attributed
- Claude Code
- GitHub Copilot
- Cursor
- Codex
- Gemini
- Devin
In practice
“We thought our bottleneck was deployment speed. It turned out to be handovers.”
FAQ
Frequently asked questions
Does Agile Analytics track AI spend and AI-assisted work?
Yes. The AI Impact page shows adoption per tool, what the tools cost and return in euros per AI-assisted merge request, and how fast AI-assisted work moves compared with the rest. It also sets change failure rate in AI-heavy repositories against AI-light ones. See the AI Impact page.
Which AI tools do you attribute?
Claude Code, GitHub Copilot, Cursor, Codex, Gemini and Devin. Cost combines the licences you pay for with metered usage, synced every night from Anthropic and OpenAI. The page shows how much of your work it could attribute, so you know how far to trust the share.
Do you rank individual developers?
No. Health scores, insights and reports are per team. Some pages name people where the work needs a name, such as who reviews whom or which repository rests on one person. That is there to spread the load, not to rank anyone.
Where is my data processed?
In the EU, under a GDPR data processing agreement. The privacy statement covers the rest.
What does it cost?
Per collaborator a month, billed yearly: €12 for Start-up, €18 for Scale-Up and €23 for Enterprise. Monthly billing costs more. There is no seat minimum, and the first 60 days are free. See pricing.
Can it record hours for a WBSO report?
Yes. The Hours page records booked hours per person and project against the contract, and exports a report. Dutch companies use the export for their WBSO, the Dutch R&D tax credit.
How long does setup take?
You connect your Git host, Jira, monitoring and Slack from the settings page. The first health report arrives the morning after. The setup guides cover each tool.
See what AI delivers, on your own data.
60 days free. Then from €12 per collaborator a month, billed yearly. No seat minimum.