Seven services, from the first check to everyday use.

Most German leadership teams are not short of AI ideas. They are short of answers to three plain questions. Are we allowed to use this? What does our AI actually cost us, and what does it return? Can our people use it safely, and can we show that they do? The seven enterprise services are built around those three questions, and each one ends in a document a board can sign.

Are we allowed to?

The legal frame changed twice in two years. The AI literacy duty has applied since February 2025. The transparency duties under Article 50 have applied since August 2026. In July 2026 the Omnibus set 2 December 2027 as the date for Annex III high-risk obligations, according to White & Case's summary. That date sits inside the next budget cycle.

The AI Act Readiness Sprint answers the first question for every system a company runs, including the ones nobody approved. The AI Governance Setup then turns the answer into an operating model: who may introduce AI, who approves it, who monitors it and who answers for it. Both are technical and organisational assessments that hand borderline cases to your counsel with the reasoning attached.

What does it cost us?

After two or three years of pilots, many large companies cannot say how much they spend on AI, let alone what it returns. Licences sit in IT, usage fees in business units, and agent features arrive inside software already paid for. The AI Estate Audit & Mapping maps all of it, with costs, usage and vendor facts, and ends with a decision for each item: scale, fix, consolidate or stop.

For the single largest bet, the AI Opportunity Case builds the business case properly. It compares real options, including doing nothing, and models full costs with the assumptions that can flip the result. A customer-service Scenario shows why that matters: a best case of +€11,336 becomes −€1,624 when each AI-written reply needs 36 more seconds of human checking.

Can our people use it safely?

People already use AI, approved or not. AI Literacy & Adoption trains them by role on the tools they actually use, writes a usage policy the works council can review, and documents everything in an evidence file. The AI Readiness & Maturity Check gives leadership one shared picture across all three questions in a single day, and the AI Governance Retainer keeps that picture current every quarter.

How the services connect

The services are designed as a path rather than a menu. A Mittelstand company typically moves from the Check to the Sprint, then to literacy and a retainer. An enterprise often starts with the Sprint or the Estate Audit, builds governance, runs an Opportunity Case for its biggest bet and keeps everything current through the retainer. The register built in one service is reused in the next. It belongs to the client throughout.

Every service draws on the same market intelligence: the AI Registry and AI Universe, which hold more than 100,000 records and thousands of dated changes, and the Good AI Prism market study, which maps 961 offerings . That is why vendor facts in an ETHORITY deliverable carry a source and a date rather than a sales claim.

Who answers for the result

Sten Franke leads the work and signs every result. Specialists join by engagement, from leadership alignment to the human oversight that has to work on a Monday morning. Prices are published, net, with 50% at order and 50% on delivery, and travel within Germany is included. The Opportunity Case is agreed in a written proposal after scoping, because its scope depends on the decision.

What the first conversation looks like

Every engagement starts with a 30-minute call. Bring the question that is actually on the table: an auditor's request, a works council letter, a vendor proposal, a budget line for 2027. Sten Franke will say which service fits, which one does not, and what you would need to prepare. Sometimes the honest answer is a smaller step, such as an AI Decision Brief for a shortlist, and the call says so.

What follows is a written scope with the package, the dates, the people needed on your side and the acceptance criteria. Those criteria are concrete: every known AI use in the register, every use case with an owner, every key assumption with a source. They are what the result is checked against before it is signed.

The common thread is accountability. Each service ends with a named owner, a decision and a date for the next review. That is the difference between an AI policy and AI that an organisation can run.