Core execution principle

First load and stabilize the master sheets in Tally and GBCFO, then proceed feature-wise polishing.

Why:

• planner trust depends on mapped item masters
• receipt trust depends on reconciled advance receipts
• stock watchlists depend on recurring extracts
• UAT quality depends on real operating data, not placeholders

Order of work

1. Freeze and complete master sheets
2. Load Tally-owned upload sheets
3. Collect Tally extracts
4. Complete shared crosswalks
5. Load GBCFO-owned setup sheets
6. Run UAT on real data
7. Polish features based on observed gaps
8. Expand release in a controlled way

Phase 0 — Freeze the contract

Goal

Lock structure, owners, cadence, and source-of-truth rules.

Required sign-off

• founder / product
• Tally team
• business / finance / ops

Exit criteria

• 3 Tally upload sheets accepted
• 2 Tally extract sheets accepted
• GBCFO-owned setup sheets accepted
• no open ownership debate remains

Phase 1 — Tally master loading

Tally upload sheets

• tally-item-master-upload.csv
• tally-vendor-master-upload.csv
• tally-item-bom-upload.csv

Tally extract sheets

• batch-stock-ledger-extract.csv
• advance-receipts-extract.csv

Outcome

• item and vendor identities are trustworthy
• BOM where needed is visible in Tally
• stock and receipt facts are available for V2 watchlists

Phase 2 — Shared crosswalk completion

Critical shared sheets

• tally-item-to-gbcfo-item-mapping.csv
• patient-advance-receipt-reconciliation.csv

Outcome

• planner can read mapped items
• receipts can be linked to active cycles
• pilot rows become trustworthy for UAT

Phase 3 — GBCFO setup loading

GBCFO-owned sheets

• gbcfo-item-enrichment.csv
• gbcfo-vendor-enrichment.csv
• gbcfo-procedure-master.csv
• gbcfo-procedure-bom.csv
• gbcfo-doctor-fee-master.csv
• gbcfo-revenue-recognition-master.csv
• gbcfo-patient-cycle-register.csv

Outcome

• rules and setup reflect the actual operating model
• V2 surfaces show meaningful rows, not placeholders

Phase 4 — UAT on real data

UAT focus

• item master and purchase planner review
• vendor mapping and enrichment
• procedure and BOM correctness
• cycle progression and milestone logging
• receipt reconciliation
• doctor fee accrual
• governance review

Exit criteria

• workflows complete without engineering hand-holding
• watchlists are understandable to business users
• blockers are specific and short

Phase 5 — Feature-wise polishing

Only after real usage

Polish in this order:

1. import UX and validation
2. mapping/reconciliation UX
3. watchlist clarity and wording
4. governance workflow support
5. planner tuning
6. deeper usability improvements

Non-goal

Do not start broad new modules before the Tally-first loop is stable.

Role-wise responsibilities

Development-side roadmap

Dev Sprint A — Import enablement

• CSV upload for GBCFO-owned sheets
• CSV upload for shared crosswalks
• row-level validation and error reporting
• recompute/refresh after import

Dev Sprint B — UAT support

• blocker fixes from live usage
• wording cleanup
• validation tightening
• mapping and reconciliation usability fixes

Dev Sprint C — Focused polish

• audit trail
• governance UX improvements
• planner explanations
• exception filters and views

Suggested 4-week cadence

Week 1

• freeze contract
• complete Tally uploads
• receive Tally extracts

Week 2

• complete crosswalks
• load GBCFO setup sheets
• validate live rows in V2 Studio

Week 3

• run structured UAT
• log blockers and friction
• begin UAT-support fixes only

Week 4

• feature-wise polish on top UAT issues
• continue controlled pilot release

End-to-end flow

flowchart TD
    A[Freeze sheet pack and ownership] --> B[Tally team completes upload sheets]
    B --> C[Tally team shares stock and receipt extracts]
    C --> D[Finance and ops complete shared crosswalks]
    D --> E[Business and ops load GBCFO-owned setup sheets]
    E --> F[Engineering validates imports and recomputes watchlists]
    F --> G[Run UAT on real mapped data]
    G --> H{Blockers are data-contract or product?}
    H -->|Data-contract| I[Fix sheets mappings and extracts]
    I --> F
    H -->|Product| J[Feature-wise polishing and bug fixes]
    J --> K[Controlled release expansion]

Final recommendation

• do not lead with feature polish
• do not reopen broad discovery
• do lead with Tally uploads, extracts, crosswalks, and setup loading
• then use real UAT to decide the next dev polish