A working tour of the platform.

A review opens to one screen with three columns. On the left, the ingested artifacts - the carrier illustration, the prior year’s template, the policy file. In the center, the populated brief with modeled scenarios and a draft recommendation. On the right, the queue of items the review still owes a human: a missing fund weighting, an unverified assumption, a draft awaiting partner sign-off.

Nothing is locked behind a wizard. The designer can step out at any point, hand the case to a colleague, and the colleague picks up mid-paragraph. Every cell carries a provenance link to the source.

A single ultra-high-net-worth client is, in practice, a household: an insured, a trust that owns the policy, a trustee with signing authority, the family office that pays the premium, the estate counsel who reads the recommendation. Interline models the relationships explicitly. Reviews coordinate across them.

When the trustee changes, the next review knows. When two policies on the same insured drift toward MEC limits, the household view shows the conflict before the partner does. When the family office requests a face-weighted IRR across four carriers, that is one query.

Every action in the workspace writes to an immutable, append-only log: the actor, the timestamp, the input hash, the output hash, the template version, the calculation parameters. The log travels with the policy file. It exports to a printer-ready certificate that looks, reads, and files like a legal document.

When the next reviewer opens the case three quarters from now, they can replay every decision: which illustration parsed, which assumption set was used, which scenario the partner chose, which version of the template was in force. The record is not assembled from screenshots. It is queried.