Sized the real serial-tracked install base on sale.order.line: ~138 units / ~136 customers across all funders (walkers 68, wheelchairs 45, power bases 7, scooters 4, +14 with no ADP device_type). Serial# is captured ~only on equipment, so it doubles as a trackable-unit marker. ADP-only gating misses ~28 units (direct_private/adp_odsp/march_of_dimes) -> bridge should key on serial, funder-agnostic. Flags two data gaps (no-device_type units; non-ADP units lacking delivery_date) and reframes the MVP open question as volume (walkers/chairs) vs margin (powered units).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Ran Step 0 against Westin prod (westin-v19 on odoo-westin). Resolved the APP/DB placeholders (DO boxes dead; migrated on-prem to odoo-dev-app), added a dated STEP 0 RESULTS section, and corrected the open questions the live inspection disproved: no stair/porch lifts in Westin ADP data; Enterprise appointment already ships native token booking; fusion_repairs contract engine not deployed; device_type is the ADP billing-code catalog taxonomy, not the install base.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Captures the maintenance-followup design exploration so it can resume from a
Tailscale-connected environment with access to Westin production:
- fusion_repairs already has a maintenance contract/reminder/booking engine to reuse
- fusion_claims (sale.order.line + adp.device.code.device_type) is the trigger source
- locked decisions: same DB, Enterprise appointment, public self-serve token booking
- Step 0 live-inspection command pack to run on Westin prod before any code
- open questions (MVP cut, revenue mechanic, tech assignment, booking route)
https://claude.ai/code/session_011wfSKQfSWhKZcm1yzSGznW
