# Brightmoor run of show: cover note

The full run sheet is in `run-of-show-brightmoor.md`. It follows the same format as the Nvalo sheets: numbered blocks with times, then a data-pack spec, pre-work, what changed, and open items.

## The shape of the session (3.5 hours)

1. What we're doing today (5 min). No AI framing. Three builds, two rules.
2. First prompt (15 min). Top-5 expenses by campus from the NetSuite GL export into a new Excel file, verified by the attendee with a pivot. Then Claude in Excel on the same workbook. Replaces the ASCII map.
3. Build 1: HTML dashboard from the income statement export (45 min). Includes Dana's tune requests (horizontal bar, filters) and a publish step.
4. Pulse check and break (20 min).
5. Build 2: month-end flux analysis and payroll-to-GL reconciliation in Excel (45 min). Ends with Claude writing its own instructions file so next month is one prompt.
6. Build 3: monthly PowerPoint pack updated from the Excel (45 min). Includes "mirror this chart for the other five campuses".
7. ASK Ladder, monthly rhythm, guardrails, optional Snowflake demo (15 min).
8. Wrap-up and Q&A (20 min).

## Assumptions made

- Start time set to 10:00 AM PT. Times shift with the actual start.
- Every provides 2 TAs in chat since Dana declined client-side volunteers.
- The synthetic data pack is described in the run sheet but not yet built. Building it is the next step once you approve the shape.
- Snowflake is demo-only, for Dana, and only if her connector is live.
- Password-protected hosting of the dashboard is treated as an IT follow-up rather than a live build step.

## Suggested next step

Approve the data-pack spec and I can generate the Excel files, the GL exports with the seeded reconciliation items, and the two PowerPoint decks.