Impact gap
Licenses without proof
The AI spend is in the budget. What it delivers is nowhere on record. When renewal comes around, anecdotes decide.
AI deep-dive diagnosis · Stage 1 of three
What your AI delivers today. What blocks it. What it can still do.
You have been paying for AI licenses for more than a year. What has measurably changed in the way your teams work? If you don't have an evidenced answer to that question, the AI Impact Check was built for you: a diagnosis across six dimensions that shows, within 2 to 4 weeks, what your current AI usage achieves, what slows it down and where there is more to be had.
2 to 4 weeks
6 assessment dimensions
Tool-neutral
Sound familiar?
Plenty of AI in the house. Little you can show for it.
Copilot is licensed, pilots are running, individual teams report successes. But when the executive board or advisory board asks what the AI investment has changed, the room goes quiet. Things are happening, that's not the problem. Nobody is measuring them.
Unregulated use
AI is already running. Just not visibly.
Employees use their own AI tools, often privately sourced and recorded nowhere. This shadow AI is not a discipline problem but a signal: the need is there, the framework is missing.
Steering gap
Every department tells a different story
Sales raves, IT warns, HR waits. Nobody has a shared picture that would allow you to prioritize.
Evidence, not gut feeling
The gap is not your special case. It is the norm.
0 %
report a measurable contribution of AI to business success. The other half do not.
Bitkom 2026
0 %
of companies have binding rules for AI use. No more than that.
Bitkom 2025
0 %
of knowledge workers already use AI in their work
Microsoft Work Trend Index 2024
0 %
of them bring their own AI tools to work
Microsoft Work Trend Index 2024
All four numbers tell the same story: AI is already in use, it just isn't steered or measured. Before you buy the next tool, you need a reliable picture of your own usage.
The solution logic
Three questions. One evidenced picture.
The AI Impact Check answers the three questions that AI decisions hinge on. The answers come from a documented measurement across six assessment dimensions, and that measurement is also what we owe you.
What is AI already doing for us?
InventoryWhere does current usage demonstrably create value, where does it trickle away, where does it run unregulated? The tools that were never officially introduced are part of that conversation.
What blocks it?
Root causesMissing rules, unclear roles, data availability, media discontinuities, trust: we name the blockers standing between license and impact, and the risk of leaving them untreated.
What can it still do for us, and at what effort?
PotentialIdentified gaps become concrete: cases per month, cost per case, an approximate savings potential in euros. No solution promises, but a clear prioritization.
Our self-commitment
We will also tell you what to switch off.
The check names where ongoing AI use demonstrably delivers nothing and should be rolled back. And our recommendation can be: no follow-up step. That, too, is a result. A diagnostic instrument guaranteed to always find a problem would not be one.
Our stance
AI competence is not a certificate. It is the way your teams work.
A training course answers whether your employees can handle the tools. The check answers whether it lands in the actual work. Two different questions, both legitimate.
Often the most effective sequence is: measure first, then train with focus, instead of training everyone and hoping. We offer enablement ourselves, from our workshop AI at Work to further training and courses. The check tells you beforehand where they will pay off.
The assessment grid
Six dimensions. A complete picture of your AI reality.
Whether AI works is rarely decided inside the tool. It is decided by how attitude, tasks, systems, enablement, rules and measurement play together. So the check covers all six, each with a survey and in-depth interviews.
AI readiness & attitude
Expectations, skepticism, the worry about being replaced: how your teams relate to AI co-determines whether it works.
ExpectationsTrustRole confidence
Task & activity profile
Where does recurring work sit, what can be automated, what can be augmented with AI, and how deep does that reach per role?
Recurring workAutomatabilityRole depth
Data, system & tool landscape
Availability, configuration, media discontinuities and the tools people really work with, beyond the officially introduced ones.
AvailabilityMedia discontinuitiesShadow AI
Enablement, roles & multipliers
Who can already do it, who carries it into the organization, and who has the mandate and time for it? Competence helps little if it sits in three places.
CompetenciesMultipliersMandate
Governance, rules & trust
Binding usage rules, co-determination, data protection, how the systems' errors are handled: the framework that turns tolerated usage into dependable usage.
Usage rulesCo-determinationData protection
Impact measurement
Which AI initiatives are already running, which of them are demonstrably effective, and which indicators show impact early?
Running initiativesEffectivenessEarly indicators
The check does not assess individuals. What is assessed is how the organization works: aggregated at team level in a GDPR-compliant way, with minimum group sizes in every analysis.
The process
From kickoff to decision paper in 2 to 4 weeks
The check is deliberately lean. The survey takes most employees around 15 minutes, key roles about an hour.
Scoping & kickoff
Scope & participationWhich teams, which questions, what data availability? Works council and data protection are at the table from the first meeting, wherever they exist.
Assessment across six dimensions
Survey & interviewsA standardized survey for breadth, interviews with key roles for depth, plus a review of existing documents and policies.
Analysis & findings
Heatmap & evidence logicEvery rating in the heatmap carries its justification: where the statement comes from, which assumption it contains, what would change the rating. That means every finding can be challenged and discussed. That is intentional.
Results presentation
90 minutesWith leadership and key roles: findings, prioritized action areas, decision paper. And if there is nothing to do, we say so.
Results
What ends up on the table
No mood picture, no interpretation by feel: you take six work products away from the check.
- Maturity heatmap across the six assessment dimensions, with evidence logic per rating
- Findings report: where AI use creates value, where it trickles away, where it runs unregulated
- Use case register: concrete gaps with cases per month, cost per case and approximate savings potential
- Rollback recommendations: where ongoing AI use demonstrably delivers nothing, with justification
- Team AI policy as a draft: tool-agnostic and ready to connect to co-determination
- Decision paper with prioritized action areas for the way forward
And after that?
The check is stage 1. What follows is your decision.
The AI Impact Check is a standalone product with an independent result, not a door opener with a predetermined outcome. If the findings show action areas, the path runs through three stages: measure, enable, integrate.
Stage 1: AI Impact Check
Measure · You are hereAn evidenced picture, heatmap, prioritized action areas, decision paper. The basis for everything that follows, or a well-founded "no action needed".
Stage 2: AI at Work
EnableWhere the check shows missing skills and missing guardrails, the performance workshop takes over: use cases from your processes, shared rules, a transfer plan.
Go to AI at WorkStage 3: AI Integration
IntegrateWhere entire work routines need rebuilding, the integration consulting continues: AI is built into the workflows, the policy takes effect, and the change is tracked against the assessment from the check.
Go to AI IntegrationNot every organization needs all three stages, and stages 2 and 3 can also be commissioned together. Depending on the findings, further building blocks from our portfolio come into play: training & courses for focused enablement, Modern Work & Microsoft 365 for the tooling side and Future of Work consulting for organizational development.
Where it fits
The entry point is the assessment. The check is the AI depth.
Both instruments measure, but not the same thing. The Organizational Effectiveness Assessment is our entry point and shows across nine assessment fields where your organization loses productivity overall. AI is one field out of nine. When that field is the actual question, the AI Impact Check goes deeper there.
Entry point · Whole organization
Organizational Effectiveness Assessment
"Where does our organization lose productivity?" Nine assessment fields across collaboration, leadership, processes and digital work. The entry point when the question is broader than AI.
Go to the assessmentDeep dive · AI-specific
AI Impact Check
"What is AI doing for us today, what blocks it, what can it still do?" Six dimensions, 2 to 4 weeks, an evidenced picture. As a deep dive after the assessment, or directly when AI is the clear question. You are here.
Our own practice
We only hold others to standards we meet ourselves.
We work AI-integrated ourselves and make our principles public: our AI policy shows the rules under which AI is used in our own work.
Common objections
Fair questions, honest answers
Does an external team know us well enough to assess this?
We do not assess your business, but a recurring pattern: how organizations adopt new ways of working. The assessment grid draws on 20 years of consulting experience in modern work and organizational development, among others from working with Atruvia and 250+ organizations in the German cooperative banking group. The outside view is part of the value: people tell externals what they do not say internally.
Doesn't our IT service provider do this?
Microsoft partners measure within the Copilot frame, which means activation rather than impact. Tenant dashboards cannot see shadow AI. We work tool-neutrally and build a picture of which tools are in circulation in your teams, including the ones never officially introduced.
What if nothing comes out of it?
Then you know that. Your next license decision has a basis in numbers, you keep the policy draft, and our recommendation can be: no follow-up step. A null finding is a legitimate result.
Do you assess our employees?
No. What is assessed is how the organization works, never individuals. Every analysis works with team values aggregated in a GDPR-compliant way, plus minimum group sizes.
We already run trainings. Isn't that enough?
A training course answers whether your employees can handle the tools. The check answers whether it lands in the actual work. Both are legitimate, and often the most effective sequence is: measure first, then train with focus. The check gives you the basis for that.
The works council will block this.
That is why co-determination is part of the product for us, not a project risk. The assessment runs on aggregated team values, the policy draft is designed to connect to co-determination, and works council and data protection are involved from the kickoff.
Our teams are already overloaded.
The survey takes most employees around 15 minutes, key roles about an hour. And the overload itself is part of what we measure: the check shows where AI can take work off people's plates.