Statement · published with the site

How we use AI

We build systems that make your judgment legible to a machine, so it would be strange not to say how machines touch ours.

In what we publish. AI assists with drafting, structuring and checking our written material. A human sets the argument, supplies the facts, verifies every figure, and is accountable for what goes out under our name. Nothing is published that a person has not read end to end.

In client work. Named in the engagement's Statement of Work before work starts. Client confidential material is never used to train a third-party model, no tool that would move your material off your machine is used without your written consent, and every output is reviewed by us. We remain responsible for the deliverables — the tools do not absorb that.

In our software. The vault kit runs inside Claude, so your prompts and the material you point it at reach Anthropic under Anthropic's terms. The vault itself stays on your disk, and nothing in it reaches us. Where material is confidential, the kit marks it and can draft a de-identification, but it never decides one for you.

The disclaimer

This block sits in the footer of every page, under the whitepaper's title page, and at the top of anything we publish. Verbatim, so it stays recognisable:

Published by Intellectual Capital Systems LLC. Written to be useful, not to be relied on: this is not legal, financial or professional advice, and reading it creates no consulting relationship. We take on work under a signed Statement of Work and not before. Figures are marked provisional until they have been checked. Views are the author's own and not those of any employer or institution he is affiliated with.