SPEC NOTES (not on-screen copy) - My Book CSV Import wizard concept v0.1 - 2026-07-30 - PEA-567.

Intercom-pattern import wizard (upload -> map columns -> preview/validate -> import -> report), adapted per PEA-567. Built as a SHARED component: nothing My Book-specific in the generic layers; second consumer = agency agent-roster upload (PEA-550 adjacency).
Interactive simulation: "Use sample book CSV" loads a fake 7-row spreadsheet (2 problem rows; 1 dedupe row that UPDATES an existing record; 1 SAME-CLIENT-TWO-PRODUCTS pair proving dedupe operates on POLICIES not people - the second product imports as a second policy linked to the same client). "Try a CSV with an SSN column" demos the hard rejection (AC: SSN/banking columns rejected, never ingested).
Column mapping: auto-detected matches tagged green AUTO; manual picks tagged MANUAL; unmapped source columns listed as visibly skipped. Required minimum: client name + carrier + date sold.
Report: created / updated / skipped tiles + downloadable error-row file (stubbed). Batch tag renders on My Book with one-action "Remove this import" undo.
POLICY NUMBER (PEA-569): now a mapping target + sample column (some rows blank, matching reality). The Rosa dedupe row demonstrates update-fills-empty: her existing record has no policy number until the import supplies it. In-app intake = PEA-569 (Mark-as-submitted modal + drill-in edit).
AGENT SPLIT (Aaron 7/30): captured per POLICY as partner name + percentage; new mapping target auto-matched from the Agent Split column (two sample rows carry values, fake names only). Renders as a Split column on book rows and per-policy sub-rows; the drill-in is its natural home in the real build. BUILD IMPLICATIONS: est. commission must reflect the agent SHARE after the split, and split data is comp-sensitive (visibility follows commission rules). Open at review: more than two agents on a split; linking the partner to a real user account.
MY BOOK VIEW OPTION (exploration, Aaron 7/30): toggle By policy (default, one row per policy) vs By client (one row per unique person; multi-policy clients aggregate as 2 carriers / 2 types with summed premium + face; click to expand per-policy sub-rows incl. policy numbers). Decide the default at review.
All data is fake. Feature ships behind a user-level flag (PEA-553); first tester = an agent with a real spreadsheet (Evan).

My Book

Your sold business. Import your existing book so everything lives in one place.

Client	Carrier	Type	Monthly	Face	Date sold	Policy #	Split	Source

Import your book

1 UPLOAD

2 MAP COLUMNS

3 PREVIEW

4 IMPORT

5 DONE