Strategy and starting point
First somebody has to know what should be built at all.
Most attempts do not fail at the build but one stage earlier: it is not clear which process justifies the effort, in what order, and cut which way. The audit answers that - using your processes and your figures, not a rule of thumb.
€2,500 · 2-3 weeks · credited against the follow-up project
The gap before the technology
The harder question is not how, but what.
When a company talks about AI, the conversation turns to tools almost immediately - which model, which subscription, which vendor. The question that comes before it is a different one: which of our processes justifies the effort at all, and what would a system have to do for something to genuinely change.
A company can rarely answer that alone, and that is no criticism. It needs two viewpoints at once: your own process with everything that has grown into it, and where these models are strong today and where they tip over. Only the company knows the first; only somebody who builds with them knows the second.
The audit brings the two together. That makes it not a measuring instrument but the part of the project where the strategy is formed - and it is deliberately written so that you can act on the result without us.
Why there is no quote at the start
The bottleneck in an AI project is almost never the build. In a mid-sized company the data sits in three systems, a spreadsheet on a network drive and in a colleague's head. An agent is built in two days - working out which process is worth it at all, and whether we can technically reach the data, is the real work.
That is why the audit comes first. It costs a manageable fixed amount, produces a document you can use without us, and is credited in full once we build. So the only risk you take is the audit itself.
What you get
- Process recording
- We look at how things run in the company and talk to the people who do the work every day - not only to management.
- Where you stand
- Which stage is realistic for you - a single process, several in sequence, or a system across departments. That decides both order and budget.
- Ranked candidates
- Three to five processes suited to automation, each with estimated effort and expected benefit.
- Data access check
- The honest answer to whether we can reach the data needed - including when it is no. This is where most projects fail, and they fail late.
- Autonomy
- For each candidate, a proposal for what the system may decide itself and what a person signs off.
- A plan with prices
- An order of work and a reliable price per step. After that, nothing is guesswork.
The document is yours. Use it internally, hand it to another provider, or put it in a drawer - there is no obligation to build with us.
How the audit runs
- 01
Kickoff
One conversation: what hurts, where time is lost, which systems are running.
- 02
Recording
We are on site, watch the processes and check the state of the data.
- 03
Analysis
Assess candidates, estimate effort, set the order.
- 04
Handover
We present the document and go through it with you.
Before that
Five questions you can answer without us.
Whether a routine carries an automation at all can be judged in two minutes - repetition, rule-governance, error cost, volume, reversibility. The self-check gives a verdict per criterion and names the one question it cannot settle.
To the self-checkFrequently asked
What does the AI process audit cost?
€2,500 plus VAT. The amount is credited in full against a follow-up project if you commission the build from us.
How long does the audit take?
Usually two to three weeks from kickoff. Most of that is scheduling with your people, not our working time.
Do we have to continue with you afterwards?
No. The document is yours and is deliberately written so another provider could work from it. There is no lock-in.
What if the audit finds that nothing is worth doing?
Then we say so. An honest no after €2,500 is considerably cheaper than a project that fails on the data after five-figure effort.
Who on our side has to invest time?
The people who run the process day to day - usually two to four of them for one to two hours each. Without them a process recording is worthless.
Do we need an AI strategy first?
No - the audit is where it is formed. A strategy written before the process recording is guesswork; afterwards it rests on your processes and your figures.
Do you only advise, or do you build as well?
Both, and with the same people. The audit is the advisory part; after it we build and take on operations. There is no handover in between to a third party who has to read themselves in again.
Tell us where the time is going.
A short conversation is enough to settle whether an audit makes sense for you at all. That costs nothing either.