You have to defend this to compliance. We wrote it so you can.
The same analysis your client receives reads their holdings against commitments they ratified themselves, with the source of each claim attached. Nothing in it is a recommendation, a rating, or a projection. The published method lives at a stable address; the disclosure block that will accompany it is not yet published — see below.
No security is recommended, rated, or ranked. The analysis says what a holding does relative to a sentence the client wrote, and stops there.
Where a judgment rests on a sourced fact, it resolves to a verbatim excerpt with its source and the date we retrieved it. You can hand a client the passage, not our summary of it.
Each report carries a version, a methodology version, and a record-state id, so a file note can cite the exact report you read. Prior states are never rewritten — but they cannot be retrieved yet, so cite the version in hand.
Advisors don’t avoid this conversation because it lacks value.
They avoid it for three reasons. Raising a client’s values can read as advocacy. They are often guessing at what the client believes. And the method behind any such analysis is the first thing a skeptical client attacks. Every report this product generates is built to remove those three exposures, whether or not the advisor commissioned it.
Every report opens with an observation about the portfolio, never a recommendation — and its disclosure states plainly that no transaction is proposed.
The commitment behind every judgment is quoted from what the client wrote, with the date it was written. Nothing is inferred from what an advisor assumes a client believes.
The rules a judgment follows are published at a stable public address, open without a sign-in — including what that page does not yet publish. Not authored by the advisor, and not private to any one firm.
Each analysis a foundation shares with its advisor carries this framing built into the report itself — the opening observation, the client’s own quoted words, the published method. Each one is generated fresh from that report’s own findings.
The disclosure block we owe your compliance desk.
A disclosure block is one short, fixed paragraph, pinned to a methodology version. It states what a report is and is not: an analysis of holdings against a client’s own stated values — not investment advice, not a recommendation to buy or sell any security, not a performance projection, and not an ESG rating. Versioned rather than freehand, it would be reviewed once per version instead of once per report. Its id would change the moment the methodology it is pinned to did, so approving the wrong version shows at a glance.
That is the design, stated in the conditional because it is not yet built. And this product never claims a block is “compliance-approved” — approval is an act a firm’s own desk performs, on the wording it actually reviewed.
The disclosure block itself, in the exact wording a compliance team would review, and its version history. The product does not yet keep a versioned disclosure-block registry, so there is no block to retrieve and no history to check against. Nothing above is citable today. Publishing a sample paragraph would create the impression of a stable, reviewable text that does not yet exist to be reviewed.
A report’s credential footer carries the methodology version current when it was saved, where one was recorded. It cannot be matched against the published method today: that page states the current rules but does not yet carry a version of its own, and names that gap itself.
A stable address a compliance desk can bookmark once.
/method states the rules a judgment follows and the vocabulary it uses: no composite score, bins for judgments and numbers for operations, ranges instead of point estimates, every claim one click from its receipt, and the five distinct kinds of missing data. A company that declines to publish and a measure that doesn’t apply to its sector are not the same finding. What each bin asserts, the evidence floor a negative judgment has to clear, and the source registry by trust tier are published there too. It requires no sign-in and carries no tracking, so a client’s compliance reader can open it cold.
The method page also states, at the foot of each section, exactly what it does not yet publish — a dated changelog of prior versions chief among them. This page follows the same rule for the disclosure-block registry above.