Paul Miles — CLM Glossary Squad Structure & Delivery Method (deck)

Tag: S-2026-06-01-paul-clm-glossary-squad Type: own-writing Author(s): Paul Miles (Red Strata) Date of source: 2026-06-01 Date ingested: 2026-06-01 Authority weight: own-writing. Raw file: PowerPoint on disk at C:\Users\pgmil\OneDrive\Documents\_ClaudeWorkspace\01 Taxonomy\CLM_Glossary_Squad_Structure_and_Delivery_Method.pptx (3 slides, black text on white, native editable diagram).

What it claims

A 3-slide executive summary for CLM programme leadership recommending an agile squad structure and delivery method to accelerate creation of the CLM business glossary across all Day 1 and Day 2 credit processes, with mastered definitions leading the imminent Day 1 system build.

Headline recommendation: stand up two empowered, cross-functional glossary squads running a one-week sprint cadence with a fast-track approval route, using AI to consolidate every business term into a single mastered definition — directly resolving the inconsistencies already found between the Matrix, the ODS and the Day 1 system design.

Sizing basis (Paul-confirmed inputs, 1 June 2026): ~500–1,000 terms total; business SMEs available ~2 days/week; Day 1 critical subset prioritised. Planning throughput: ~120–140 terms/week per squad → ~250/week combined once AI clustering is live; Day 1 critical terms mastered in weeks 1–4 (4–6 weeks); full baseline glossary in ~5–6 weeks; flex to a 3rd surge squad only if scope nears the 1,000+ end.

Squad composition (per squad): Squad Lead / Product Owner ×1; Business Data SMEs / Definition Owners ×2–3 (empowered to make and recommend definitions, ~2 days/wk); Data / Business Analysts ×2; Data Steward / Governance Lead ×1; plus shared-across-squads Agile Facilitator (~0.5), AI / Data Engineer (~0.5) and on-call Regulatory / Risk SME (~0.3). Core dedicated headcount ≈ 6–7 per squad, with the facilitator, AI engineer and regulatory SME shared to protect scarce SME time. Squad A works the Day 1 critical path; Squad B works Day 2 (ODS, monitoring, servicing, collections) in parallel.

Fast-track governance: squad SMEs empowered to approve definitions in-squad (SLA 3–5 working days); AI-identified unique terms cleared by consent-based bulk approval; only unresolved conflicts escalate to a standing fortnightly slot at the existing CLM Design Sub-Committee. Leverages existing programme and BAU fora (DQ Forum, Enterprise DG Council, BCBS 239 attestation) rather than creating new bodies.

Delivery-method flow (the editable diagram): (1) Harvest terms from Matrix, ODS, Day 1 design, Policy & Covenants → (2) AI cluster & de-duplicate, splitting unique vs conflicting → (3a) unique terms to fast-track bulk approval / (3b) conflicts flagged with an AI-proposed consolidated definition + rationale → (4) SME workshop decides (escalate only if unresolved) → (5) master definition with Policy / Process / IRB / BCBS 239 tags → (6) publish to the governed master (Excel → SharePoint → Confluence / Comala).

AI is used to: harvest and cluster terms across sources; identify unique values for fast bulk approval to build the baseline quickly; detect conflicts and propose consolidated definitions with rationale; suggest Policy/Process/Regulatory tags; and draft/QA via M365 Copilot — humans decide, AI accelerates.

Notable quotes

• “Humans decide; AI accelerates.” (slide 3 AI panel)
• “Definition of Done: single mastered definition + complete business metadata + fast-track approval + published to the governed master.” (slide 3)

What’s speculative vs. asserted

• Asserted (Paul’s design intent): two-squad structure, role mix, one-week sprint cadence, fast-track governance route, AI usage points, and the Excel/SharePoint/Copilot + Confluence/Comala tooling pattern.
• Speculative / planning assumptions: throughput figures (~120–140 terms/week/squad, ~250 combined) and the ~5–6 week baseline timeline are a planning model, not measured — flagged on the deck as “re-baseline once the harvested term count is confirmed”. The 500–1,000 term scope, ~2 day/week SME availability, and 4–6 week Day 1 deadline are user-stated inputs, not yet validated against a harvested inventory.

Topics this feeds

• CLM Glossary Acceleration Squad — the project page synthesising this design.
• Relates to CLM Data Taxonomy Governance (Confluence/Comala governance pattern) and the CLM Transformation Programme.

Open questions raised

• Actual harvested term count across Day 1 + Day 2 (drives squad count and timeline).
• Size of the Day 1 critical subset that must lead the system build.
• Whether SME availability holds at ~2 days/week once squads are live.
• Whether the CLM Design Sub-Committee will host the standing fortnightly glossary-decision slot or a separate fast-track authority is needed.