Our core is People Helping People

AI Program Scale-Out

Capability Mapping and Proposed Discovery Session

Prepared for Fibre Converters, Inc.|July 9, 2026|Prepared by 'corePHP'

Current State

FCI AI & Data Program

FCI has built the Data Universe, a self-contained data and knowledge platform for AI coding agents, with a portfolio of 13-plus analytics projects on top of it. The platform currently provides:

  • A single source of truth: canonical truth files and a structured data catalog of connections, tables, metrics, and business rules.
  • Guard-railed, read-only access across GP Dynamics, BVP Oracle, OMP Oracle, Ignition, and QuoteIt through a shared code library.
  • Centralized secrets, full version control, connectivity self-tests, and automated encrypted backups.
  • Institutional memory: a decisions and learnings log plus per-project truth files, so a session can resume without context loss.

Stated objectiveMove the platform from individual technical use to organization-wide use. Per David Posey, the core is in place; scaling it across the organization is the open problem.

Capability Mapping

Requirements for organization-wide use, mapped to 'corePHP' components

Requirement to scale 'corePHP' component Status
Access for non-technical staff Guided, no-terminal workspace over the existing data, so staff without coding skills can query it and generate outputs. Available
Shared, multi-user memory Team knowledge store that captures, de-duplicates, and retrieves decisions and learnings, loaded into each session automatically. Available
Multi-project orchestration A board that runs multiple agent tasks concurrently, each isolated and tracked from request to completion. Available
Governance over production access A mandatory review step, secrets scoped away from the agent, and per-project boundaries. Available
Current system and API context for agents Self-hosted documentation service that serves up-to-date references to agents. Available
Role-based access model Access and role layer over the shared store, configured to FCI departments and audit requirements. Custom build

Open Questions

Items to resolve during discovery

  • Tooling. FCI's agents run on Cursor. The orchestration and shared-memory layer would integrate with, adopt, or run alongside that setup. This is the primary integration decision.
  • Workload mix. The ratio of interactive build work to scheduled data jobs (daily briefing, forecasts, month-end audit). The two scale differently and are sized separately.
  • Access and governance. Department-level access requirements, and the roles and audit standard required for organization-wide deployment.

Engagement Options

Engagement models

Option A

Discovery Sprint

1 to 2 weeks

Target architecture, the tooling integration decision, and a phased roadmap with estimates.

Option B

Pilot

4 to 6 weeks

Shared-memory layer plus one organization-facing workflow (for example a self-serve finance view) deployed to a small group.

Option C

Org Rollout

Phased

Access model, governance, training, and department-by-department scale-out.

Next Step

Proposed discovery session (30 minutes)

  1. Current platform review10 min
    Walk through the Data Universe and confirm the scaling requirements.
  2. Component demonstration10 min
    Review of the shared-memory layer, orchestration board, and self-serve workspace.
  3. Fit and options10 min
    Confirm the tooling and workload approach and select an engagement model.