01Who we are
Probative Co is an independent compliance and AI governance practice working with UK law firms. We run independent AML file reviews, build DPIA and AI policy packs, prepare firms for ISO/IEC 42001, and design private AI deployments. We are not a law firm and we do not sell software.
The index began as an internal document. The same questions arrived from COLPs and MLROs in every engagement — where does the data go, does privilege survive, can we evidence what it did — and there was nowhere independent to point them. So we published the document.
02How it is funded
- No vendor payments. No sponsorship, affiliate links, paid placement or featured slots. We have declined all of it to date.
- No vendor previews. Scores are not shared before publication.
- Funded by client work. The index is written from the same desk that delivers engagements. It costs us delivery time and earns us the position of having looked — which is the only honest thing we can sell.
- No advertising, no tracking pixels. If you see a number on this site that looks like a metric, it is a number we typed.
03The disclosure that matters
We sell a vendor diligence pack to AI companies selling into law firms, and we advise firms on tool selection as part of audits. That is a commercial relationship with the same market we review, and pretending otherwise would be worse than the conflict itself.
What we do about it: vendors do not see scores early, do not get to influence entries, and do not get scores withdrawn by buying something. Firms paid for an audit get an honest assessment of every tool we look at, including tools whose vendors we have worked with — and we have published ordinary scores with caveats for exactly that situation. If that arrangement is not independent enough for you, treat the index as one input rather than the verdict. That is what we would do.
04Who writes the reviews
The same people who deliver engagements: a managing director with a decade building AI inside law firms, an AML associate who is a former MLCO, and a CIPP/E privacy associate. Where an entry depends on technical testing we say who ran it and on what. We do not use ghostwriters and we do not publish AI-generated reviews — which should not need saying in 2026 and does.
05Corrections policy
Any vendor or firm can challenge a score with evidence. If it holds up we correct the entry, publish the correction with its reasoning, and note what changed. If our method was wrong rather than the entry, we change the method and say so in the version history. Silent edits would make the whole document worthless, so there are none.
06Contact
Index correspondence: index@probativeco.com. Everything else: diin@probativeco.com. Firms wanting the index applied to their own stack should start with the AI use audit.