Why We Refund Below Eighty

A dossier that is not genuinely useful should not cost you anything.

That is the logic behind the automatic refund. When the confidence score on a finished dossier lands below 80, the system marks the result Low Signal and attaches a banner to the cover page. And when our system determines a dossier could not be genuinely useful, the credit comes back within the hour. No support ticket. No conversation.

We built this because we believe the advisor should only pay for work they can use.

What the confidence score measures

The confidence score on a Wealth Recon dossier is not a measure of the research quality. It is a measure of something narrower: how certain are we that the subject of this dossier is the actual person the advisor asked about, and not a same-name collision with a different life?

Name disambiguation is genuinely hard. There are thousands of people named John Smith in the United States. The score reflects the cumulative probability across disambiguation signals: geographic match, employer match, verified public role, regulatory record match, property-of-record confirmation, and professional network corroboration. Each factor lifts or suppresses the score. The final number summarises the chain.

A score of 87 means the identification is solid. The sourcing beneath it is real, and the advisor can use the briefing with confidence in a meeting.

A score of 72 means we found material on someone with this name, but one or more anchors that should confirm identity did not resolve cleanly. The dossier ships, because the research itself may still be useful, but it ships with a prominent warning; if the result also fails the usefulness floor, the credit refunds automatically within the hour.

Below 60, the dossier carries a Weak Signal designation. Everything in it should be treated as preliminary until the advisor can confirm identity through an independent channel.

Why 80 is the threshold

Eighty is not arbitrary. It sits at the boundary between a result a well-prepared advisor could confidently brief from and a result that requires the advisor to verify the subject before trusting any specific claim.

Below 80, the disambiguation pass did not find enough anchors to stand behind the result without qualification. The advisor's preparation time is not recovered by reading material that could not be made useful. The honest response is the refund.

We chose to make the refund automatic rather than manual because a credit refund should not require the advisor to file anything. The refund is the product acknowledging a limit, not a customer-service gesture.

What happens to the dossier

The Low Signal dossier remains in the advisor's research history. The cover page carries a banner:

LOW SIGNAL · CONFIDENCE 72 · SUBJECT IDENTITY NOT CONFIRMED

Every section is still readable. The source manifest is still present. If the advisor later confirms the subject through independent means and wants a fresh build, they can request a new dossier. The new run costs one credit; it triggers another refund only if the result again fails the usefulness floor.

The floor this sets

The refund is a floor on the usefulness of the product. It commits us to something: we will not deliver a briefing that could not be made genuinely useful and call it done. If the dossier does not clear that floor, the cost does not clear either.

That is the standard we hold ourselves to, and the one we believe every research service in this category should hold.

Confidence bands: Strong (90 to 100), Strong (80 to 89), Medium (70 to 79), Weak (below 70). The Low Signal banner appears at any score below 80; the automatic refund triggers when a dossier fails the usefulness floor.