# Goal — gig-RAG (corpus microsite)

**What this is:** the single review surface for the GIG corpora that feed GG — the
travel / medical / motor + all-lines knowledgebase. One page that answers, for
Anthony and for Sidd: *what is in it, how it is stored, how it is used, how complete
it is, what is missing, and what it can be used for.*

**Who it serves:**
- **Anthony** — provenance + completeness at a glance before any claim is built on the corpus.
- **Sidd (GIG)** — proof the GG knowledge layer is real, measured, and governed — not vapour.

**Success looks like:** every corpus claim on a client-facing deliverable traces back
to a row on this page; the completion scores are measured (not estimated); and the
named gaps each have an owner and a next action.

**Current open question:** does GIG supply first-party book data (premiums, nationality,
vehicle, LTV)? Without it the targeting hypothesis can't be tested — it is the single
biggest gap, and no amount of competitor-corpus depth substitutes for it.

**Goal metric:** portfolio completion ~55% → 75% by closing the 4 thin motor entities,
extending MedBench to the 9 missing insurers, and adding GCC markets.

**Last updated:** 2026-06-19