// 01 Intake
// What you should know
Eight questions we answer
What is really holding us back?
For every data layer your use case needs, we measure what limits it. From the article master to the agent log. The result is your bottleneck, quantified and evidenced, with the price of the gap.
Why does our pilot never go into production?
Because the pilot runs on curated data and operations run on your real system landscape. The audit shows which data layer blocks the jump and what it takes to turn the demo into a system that holds.
Can AI agents work with our data at all?
That is exactly what we check systematically: are there interfaces, are they documented, are the data models accessible in machine-readable form? This is the question behind every AI ambition, and it can be measured.
What do we have to do and what does it cost us?
You get a prioritised action plan with effort estimates and the expected effect of every measure. Described so that you can run it with us or without us.
How do we know the result is right?
Every rating needs a verifiable artefact from your systems. Without evidence there is no better rating, whatever is reported. On top of that: an open measurement standard, a four-eyes review before any result, and an appendix that makes every number retraceable for third parties.
What happens to our data?
It stays where it is. We work with reports, samples and logs, and the collection happens inside your systems. An NDA before any system contact, read access minimal and time-boxed, comparison values anonymised only.
Do we have to clean up our data first?
No. The audit rates the current state, that is what it is for. Cleaning up first would even distort the result. What to do afterwards is in the action plan: only the layers that limit your use case.
And if it turns out we are not ready yet?
Then you know it after three weeks instead of after twelve months of a pilot without production. The audit has an open outcome, with three possible results and a clear next step. All results belong to you, with no obligation to buy more.
// What is included
Four results and one report,
Four results and one report,
always to the same standard
Every audit runs the same way and delivers the same building blocks. You know upfront what will be on the table at the end.
// 02 Deep check
Data quality profiling, interface analysis and logs inside your systems
Phase 2
// 03 Findings report
Management summary, reliability score, bottleneck, action plan, positioning
The report
// 04 Four-eyes review
No result leaves the house unchecked
Quality assurance
// 05 Findings session
The findings with your decision makers, including an honest recommendation
Even when it says “no follow-up project”
// How it runs
Three weeks, a fixed end date
Phase 1 · Intake
A structured questionnaire with an onboarding session, then the intake dialogue: system map, ownership, access paths. Still without system access. Afterwards both sides know exactly what will be checked and which access is needed.
Phase 2 · Deep check
We read your systems, not your folders. Data quality profiling on real stock, analysis of interfaces and data models, evaluation of pipeline and agent logs. Read-only, minimally invasive, reproducible.
Phase 3 · Findings
Rating against the open measurement standard, four-eyes review, findings report and a findings session with the sponsor and the CFO.
// The report
Not a slide deck. A finding.
Every audit delivers the same report: a one-page management summary, your system landscape, the data model with measured field completeness, the bottleneck with evidence and the action plan with the expected effect. In the appendix every number is written so third parties can retrace it.
See the example report (PDF) Fictional example · ACME Inc. · real method
// Whitepaper
Is your data good enough for your AI plans? How our audit makes it measurable.
Download the whitepaper (PDF)- Why the bottleneck is almost never the model. And the numbers behind it.
- What an honest assessment covers: evidence instead of self-reporting
- The process, the four results and the three possible outcomes
// Effort & terms
What it costs you. In time and in money.
Your effort
- One half-day workshop with decision makers and the business unit
- One subject matter contact, two to three hours per week
- Read access following the checklist from phase 1. No preparation and no clean-up beforehand.
Terms
- A fixed end date
- All results belong to you
- Pricing and terms discussed in person
Request it now
An answer within 24 hours. From a person, not from a funnel.
// Three possible outcomes
The audit has an open outcome. Every result has a next step.
We close exactly the gap at the bottleneck. In the vertical slice sprint.
No data project needed. You start further up.
We tell you before you invest.
An NDA before any system contact
A confidentiality agreement before the first look. A data processing agreement wherever personal data is touched.
Evidence instead of raw data copies
We work with reports, samples and logs. The collection happens inside your systems.
Minimal, time-boxed read access
Tied to named people, limited to the relevant systems and revoked when the project ends.
Comparison values anonymised only
Our benchmark takes anonymised, aggregated maturity scores only. Never business data.