Five ways to start, from €195 net.

AI decisions come in different sizes. A team choosing between three customer-service assistants faces a different question from a board weighing a platform commitment. Treating both the same wastes time on the first and takes the second too lightly. The five entry offers are built to match the effort to the decision.

Compare: the AI Decision Brief

The smallest question is often the most common: which of these three offerings fits us? The AI Decision Brief is requested directly from Compare in the Good AI Prism. It puts the facts that decide a purchase side by side, each with its evidence class and read date, applies the pricing to your volumes and lists what to ask each provider before signing. Where a provider does not publish a fact, the Brief turns it into a question for that provider. Facts age fast: research for the Good AI Prism study found 34 product renames within a few weeks.

Plan: the AI Business Blueprint

Sometimes the question comes before any shortlist: where could AI change this part of our work at all? The AI Business Blueprint starts from one business objective and the work behind it. It maps the work, the opportunities and the technology patterns, frames the economics with every figure marked as known or still required, and ends with two to four options and a 30/60/90-day roadmap. Its one-page executive decision is written to be forwarded.

Judge: the AI Trust Record

Many AI systems are already in use and growing quietly. The AI Trust Record asks whether one named system can be used for its purpose with clear responsibility, adequate evidence, known limits and an agreed review path. It answers in ten working days with one of four decisions, pilot, redesign, expand or stop, and a Trust Chain that shows where the evidence is strong and where it is weak. A 90-day list of review triggers keeps the decision alive.

Test: the AI Trust Pilot

When the answer is "pilot", the AI Trust Pilot runs one agreed use case in one department for six weeks. The difference from an ordinary pilot is that success and stop conditions are agreed before the first live case, and the baseline is measured before AI is involved. The pilot ends with a decision package.

Decide: the AI Opportunity Case

The largest decisions, investments of roughly €200,000 or more, go to the AI Opportunity Case. It compares three to five real options, always including doing nothing, and models full costs with the assumptions that can break them. A customer-service Scenario shows the point: a best case of +€11,336 becomes −€1,624 when each AI-written reply needs 36 more seconds of checking. The result is a signed decision record, and the price is set after scoping.

How they connect

The offers lead into each other. A Blueprint often produces a shortlist for a Decision Brief. A Trust Record that says pilot leads into a Trust Pilot. An Opportunity Case that says "test first" hands its pilot plan to the same structure. When a question grows beyond one decision, the founder-led enterprise services take over with a whole-organisation view.

All five share three rules. Facts carry a source and a date. Missing evidence becomes a question or a condition. And a named person, Sten Franke, confirms the scope before work starts and stands behind the result. That turns even a small entry offer into a decision you can defend.

Choosing in practice

If you are unsure where to start, ask what you already hold. A shortlist points to the Brief. An objective without a shortlist points to the Blueprint. A system already in use points to the Trust Record. A use case that someone wants to scale points to the Trust Pilot. A large budget request points to the Opportunity Case. And if the honest answer is all of the above, across the company, the one-day Readiness & Maturity Check sorts the questions before any of them is bought.

Each offer page shows what you receive, how the work runs, what you need to bring and where the boundary lies. Each one is a decision with its evidence attached, sized to the question in front of you.