AI Usage Audit
In 1–2 weeks I check whether the AI you pay for is the best available mix of models, tools and architecture — and where the same quality costs less. The result is a list of decisions with numbers, not a deck.
Who it’s for: For companies that already run AI in production or daily work and pay real invoices for it: APIs, licences, tools.
Not for companies before their first deployment — that’s where the Opportunity Map starts. Not a compliance audit or a paper for the board either.
The setup from launch day. Today’s invoices.
The model and architecture picked at the first deployment sit untouched while the market has moved several times — and the monthly bill keeps growing. Nobody can say whether it’s money well spent, because nobody has time to measure it. The audit measures: where the same quality costs less, where you pay for unused tools, and where you use AI too little.
What you get
- An inventory of every place you use AI, with the real cost of each
- Quality and cost of your current setup measured on your real cases
- Comparison against the best available alternatives: models, prompts, architecture
- A list of swaps that hold quality and cut cost — with a percentage next to each
- Places where you underuse AI or pay for tools nobody touches
- A call for every line item: change, keep or switch off
How it runs
What you need on your side
The audit works if you have:
Pricing
€4,500–8,000 net
- fixed price for a fixed scope — confirmed on the 15-minute call
- the outcome is independent: the recommendation may be “keep it as is”
- implementing the changes — on your own from the plan, or in a sprint, priced separately
As of: August 2026
Run your own numbers.
Take what AI costs you per month and how much of it can be recovered at the same quality — then compare it with the price of the audit.
The math
how many times the audit pays back in 12 months
4.5× — how many times the audit pays back in 12 months- saving on AI costs per month
- €3,000
- saving over 12 months
- €36,000
- audit price (top of the range)
- €8,000
- the audit price pays back after
- ~2.7 mo
An estimate built on the numbers you enter — not a promise. The audit ends with measurements on your processes, and the recommendation may well be: “keep it as is”.
Before you book a call.
How long does the audit take and what does it cost?
1–2 weeks and €4,500–8,000 net — a fixed price for a fixed scope. I check whether the AI you pay for is the best available mix of models, tools and architecture, and where the same quality costs less.
What do I get at the end?
A report and a change plan with numbers: a list of decisions with their estimated effect, not a presentation. You can implement the changes yourself from the plan, or have them done in a sprint, which is quoted separately.
Is this a compliance audit or a paper for the board?
No. It is a technical and cost audit: models, tools, architecture and invoices. If you need a document certifying regulatory compliance, that is a different service and a different supplier.
What if our choices turn out to be good?
Then the recommendation is “keep it as is”, and that is also a result worth its price: you know you are not overpaying, and you stop revisiting the question every quarter.
We don’t run AI in production yet — where do we start?
With the AI Opportunity Map. An audit only makes sense once you pay real invoices for APIs, licences or tools — otherwise there is nothing to compare and nothing to bring down.
What’s next
You implement the changes yourselves from the plan — or we take the top item into an Agentic Workflow Sprint, priced separately.