# Execution Backlog_ Founder CRM and Platform Planning

**Date:** 2026-04-18  
**Status:** Draft  
**Owner:** Sagnik  
**Reviewers:** Sayan, Sourik  
**Scope:** Sequenced founder workstreams for CRM, client graph, imports, infra truth, and documentation convergence  
**Purpose:** Translate the founder implementation blueprint into an execution-ready planning backlog with ownership, sequencing, and approval gates.  
**Decision Boundary:** This backlog governs planning deliverables and future implementation preparation. It does not itself authorize coding.

## 1. Purpose

This backlog governs the founder-side convergence work required before Project Velocity can move into a clean CRM-centered implementation phase.

## 2. Delivery Principles

- reconcile current truth before adding new architecture
- define canonical data before building new UI
- preserve raw evidence before flattening into operator views
- keep Oracle adjacent to CRM, not above it
- respect owner boundaries across Founder, Sayan, and Sourik

## 3. Workstream Backlog

### Workstream A: Repo and Sprint truth reconciliation

- objective: align all founder planning to the current repo rather than outdated assumptions
- tasks:
  - reconcile Sprint 1 fact table against repo and handoffs
  - identify stale vs current claims
  - mark legacy feeders and current owner boundaries
- definition of done:
  - reconciliation matrix completed and usable as a review artifact

### Workstream B: Infra and operator truth

- objective: create one founder-safe deployment and routing reference
- tasks:
  - consolidate Linux origin truth
  - consolidate ingress truth
  - consolidate Comfy/GPU/NVMe truth
  - define endpoint ownership matrix
- definition of done:
  - operator can understand the runtime topology without opening multiple older docs

### Workstream C: CRM product architecture

- objective: define CRM as a top-level product module
- tasks:
  - define CRM IA
  - define Oracle adjacency
  - define contact/account/opportunity/task/import/client-360 surfaces
  - define reuse strategy for existing Oracle pipeline/detail components
- definition of done:
  - CRM product shell is clear enough for frontend implementation planning

### Workstream D: Canonical domains and root API

- objective: define the target persistence and route model
- tasks:
  - define `crm_*`, `intel_*`, `inventory_*`, `oracle_*`, `core_*`, `workflow_*`
  - map current tables into target domains
  - define root route families and ownership
- definition of done:
  - DB/root API spec is approval-ready

### Workstream E: Import and stewardship model

- objective: define the raw-to-canonical import architecture
- tasks:
  - define CSV-first ingest pipeline
  - define source profile and mapping manifest model
  - define AI assistance and approval gates
- definition of done:
  - import adapter and schema blueprint are implementable

### Workstream F: Synthetic data generation prep

- objective: prepare a coding-agent swarm brief for realistic synthetic client graphs
- tasks:
  - define canonical data requirements
  - define property pool and client personas
  - define deliverable formats
- definition of done:
  - a swarm can generate 250 full synthetic client graphs without further architecture decisions

### Workstream G: Monolithic truth consolidation

- objective: create the founder-owned single-reference artifact
- tasks:
  - define chapter order
  - seed initial chapters
  - cross-link founder pack documents into the monolith
- definition of done:
  - a new engineer can navigate the product architecture by chapter

## 4. Acceptance Gates

### Gate 1: Current truth locked

- reconciliation matrix completed
- infra/operator truth completed

### Gate 2: CRM and DB architecture locked

- PRD, SRS, blueprint, adapter, and DB spec aligned

### Gate 3: Execution readiness locked

- backlog, dependency matrix, and synthetic-data brief completed

### Gate 4: Founder reference locked

- monolithic truth artifact created and seeded

## 5. Ownership Model

- Founder
  - CRM
  - import architecture
  - client graph
  - monolithic truth
- Sayan
  - Oracle template DB
  - inventory and multi-surface implications
  - admin/control implications
- Sourik
  - runtime route-family and orchestration implications

## 6. Sales Readiness Criteria

- contact list and client 360 story is explainable
- Oracle’s role relative to CRM is clear
- import story is believable for existing CRM users
- infra/operator deployment truth is clean enough for buyer confidence

## 7. Risks

- route-family sprawl without canonical domains
- overloading Oracle with CRM responsibilities
- designing a perfect AI graph that operators cannot practically use
- under-planning migration from current tables

## 8. Immediate Next Actions

- finalize all founder pack documents
- review with Sayan and Sourik
- lock canonical domain naming
- approve the synthetic data brief
- use the monolith as the next planning anchor

## 9. Bottom Line

The founder backlog is not about adding more product ideas. It is about forcing the existing product into one coherent operating model before more code compounds the inconsistency.