One AI decision, tested against your own data. In writing, within 5 working days.

A fixed fee, agreed before we start, that never moves with the verdict. It is a small fraction of the proposal it tests; if it wouldn’t be, we’ll tell you on the call. You get the number on the first call, once we know which systems and decisions are in scope.

We are new, and we won’t pretend otherwise. That is why the method is published, including the cap that stops strong pillars hiding a weak one. Every finding points to records your own team can open and check, and the scope line says what we didn’t check. You can stop at the memo. Our first clients get a founding rate in exchange for a private reference call once they’re happy with the work, and only with their consent: we never name a client, fund or target publicly without written consent.

Standard exports (spreadsheets downloaded from each system) from any two or more systems the decision depends on: CRM, billing, ERP and finance, HR and payroll, projects, support. Spreadsheets count. We send the exact field list first, so one person on your team can run the exports. If there is a proposal, plan or business case, send that too.

A Truth Score Read is delivered within 5 working days of complete exports. An AI Investment Review takes about 2 weeks. AI Value Tracking runs every quarter. If an export is missing fields we need, we say so straight away; the clock starts when complete exports arrive.

No. One person on your team runs standard exports from the systems behind the decision and sends them under NDA. We send the exact field list first. You never log in to anything and nothing is installed.

It arrives under NDA and our data processing agreement. The full list of sub-processors comes with our data processing agreement, before any data moves. Any transfer outside the EU is covered by Standard Contractual Clauses. Raw exports are deleted after the engagement.

The Truth Score is rule-based: the same data, scored on the same date, gives the same score, and no AI model decides it. We do use a language model (Anthropic’s Claude, through its commercial API) to read the plan or proposal you send and to help draft the memo’s wording from our findings. A person checks every line before it reaches you. If you’d rather we didn’t use it on your engagement, say so and we do that part by hand.

Yes. Send it with your exports. We take each claim it makes, work out the data that claim depends on, and check whether that data is there and can be trusted. Not whether the AI is clever, but whether your data can feed it. The verdict is plain: sign, fix first, or walk away.

Because by phase one you have signed. Their data phase is scoped, priced and staffed by the firm that will run the phases after it, so it is designed to lead to them. Ours comes before the signature, from a firm paid the same whatever the answer, and it tests whether the proposal’s own assumptions hold in your records. Use the memo to re-scope or negotiate the proposal.

Then the memo says sign, in writing, and you go ahead with evidence instead of a vendor’s word. If you don’t need anything more from us, we’ll tell you that too.

The facts in it. Every number points to records you can open, and the scope line says what we didn’t check. The decision stays yours: the memo says what your data can and can’t support, and what closing each gap takes. If your team disputes a finding, we re-check it against the records with them.

No. It stands on its own. You can stop at the memo, hand the fix list to your own team, or ask for an AI Investment Review when the question becomes every AI bet in the business. You decide, and nothing is bundled.