# Your team works with AI. Can you show what it returns?

Audit · 4 weeks · fixed price

The licences are paid for, the usage is unclear, and in the next budget round someone will ask for numbers. In four weeks I measure what actually changed and what didn't.

**Next audit slot: September 2026**

- Book an intro call (20 min): https://calendar.app.google/jpSRhy2ekZWZEv3W7
- Email: business@jomaendle.com

## How you recognise it

- **The licence cost is in the budget. The return is not.** You pay per developer per month. Nothing on the other side of that line is written down.
- **The reports from your teams contradict each other.** Two teams are delighted, three stopped using it after the rollout week. Nobody knows why.
- **The numbers you already have don't carry the decision.** Acceptance rates and usage quotas say nothing about lead time, review effort or failure rate in production.

This is not a tooling problem. It is a measurement problem, and it can be solved without rating individual developers.

## The audit

Four weeks, two sessions a week, one written result.

### Week 1 · Baseline

Analysis of Git, CI and the ticket system over the last twelve months. Lead time from commit to production, review duration, change rate after merge, rollbacks. The state before AI arrived, from data you already hold.

### Week 2 · Current practice

Where agents are genuinely used today. Six to eight conversations of 30 minutes, across teams and seniority levels. These are conversations and not an anonymous survey.

### Week 3 · Impact analysis

The same metrics after the rollout, set against the baseline. Split by team, codebase age and type of work. This is usually where it becomes clear that the effect is not evenly distributed.

### Week 4 · Recommendation

Where to expand, where to stop, what the blocker is. With effort and expected effect for each measure.

## What you have at the end

- A metric set your team can keep running without me: definitions, queries, scripts
- Baseline and current value, evidenced from system data rather than self-reporting
- One page for the executive board that can carry a budget decision
- Three to five prioritised measures with an effort estimate
- All raw data and analysis scripts, in your repository

The results come as a document and as scripts, not as a slide deck.

## Why this measurement holds up

### Measurement without performance monitoring

Tools built in the US measure per developer. In Germany that triggers works council codetermination, and in many companies it is the reason a rollout stalls there.

I measure at team and repository level. No personal analysis, no ranking, no individual profile. The metric set is built to hold up in a works agreement. I supply the description you need for it.

### I have introduced this myself, not only assessed it

As a Principal Solution Architect in a corporate group with several hundred developers, under the same constraints you know: works council, procurement, mature codebases, teams with deadlines.

## Scope and price

**Fixed price from €10,000 · 4 weeks · remote, one day on site optional**

- **Included**: Data analysis, six to eight interviews, written result, closing presentation, handover of the scripts.
- **Not included**: Implementing the measures. That is a separate mandate and you decide afterwards whether it runs with me or without.
- **Fits if**: 50 to 800 developers, AI tools in use for at least three months, Git and CI with a history that can be queried.
- **Does not fit if**: You are still before the rollout. Then the baseline is missing and the audit measures nothing. It becomes worthwhile once the tools have been in use for a quarter.

## Common questions

### What if the result comes out negative?

Then that is what the document says. An audit whose result is settled in advance is worth nothing. In that case you have solid grounds to reduce licences. That pays off faster than any optimisation.

### Do you get access to our code?

No. I need metadata: commit timestamps, pull request history, CI runs, tickets. No source code leaves your company, and that is in the contract.

### How much of our time does this cost?

Six to eight interviews of 30 minutes, one technical access in week 1, one closing session. Under ten hours on your side in total.

### Why not one of the off-the-shelf tools?

You can do that. After the audit you know which metrics are meaningful in your company. Before it, you buy a dashboard and only afterwards find out which metrics you needed.

## Next step

20 minutes. I ask the questions, you decide afterwards.

In the call we work out whether your data even carries an audit. If it doesn't, I say so in the call and not after you have signed.

- Pick a time: https://calendar.app.google/jpSRhy2ekZWZEv3W7
- Email: business@jomaendle.com
- HTML version: https://www.jomaendle.com/ai-impact
- German version: https://www.jomaendle.com/ki-wirkung