IGA Toolkit Build Plan — Structured Approach for Claude-Assisted Build
Tag: S-2026-07-17-iga-toolkit-build-plan
Type: own-writing
Author(s): Paul Miles with Claude (Cowork session)
Date of source: 2026-07-17
Date ingested: 2026-07-17
Authority weight: high — Paul’s own planning document, scope confirmed by Paul in-session
Raw file: S-2026-07-17-iga-toolkit-build-plan.md (working copies: _ClaudeWorkspace\04 Ai and Data services\IGA_Toolkit_Build_Plan_2026-07-17.docx / .md; companion record IGA_Toolkit_Build_Work_Breakdown_2026-07-17.md)
What it claims
A structured approach for using Claude to build the missing Independent Governance Assurance tool sets — the templates that make fixed-fee reviews viable. Scope (confirmed by Paul 2026-07-17): the core IGA toolkit of six tools plus the differentiating offer one-pager; Phase 1 universal tools and the ~30 AI assurance pathway artefacts are explicitly out of scope. The plan specifies, per tool: purpose, the client problem it solves, the service supported, expected output and value, and recommended format (Word for narrative/client-facing documents; Excel for scored/tracked instruments; findings register and management-response tracker combined into one linked workbook). Build sequence is three waves aligned to the 6-month transition plan: Wave 1 (Aug 2026) — findings & actions register, ToR template, assurance report template, offer one-pager; Wave 2 (Sep 2026) — interview guide set and document review checklist; Wave 3 (Nov/Dec 2026) — v2 after pilot plus an ISO/IEC 42006-aligned methodology write-up. Each tool has a ready-to-run Claude build brief with per-tool QA checks, under a universal 8-point QA gate (alignment, citation integrity, client-ready, solo-operator, fixed-fee, independence/confidentiality, technical, adversarial read). Market-demand grounding is drawn from the sourced statistics on AI Governance Maturity Gap plus external anchors (IIA AI Auditing Framework, ISO/IEC 42001/42006).
Notable quotes
- “Build 1 first — it defines the risk-rating scale and governance-dimension list that tools 3, 5 and 6 must reuse verbatim.” (Work breakdown, dependency rule)
- “Citation integrity … no invented citations. This is non-negotiable for an assurance brand.” (§6, universal QA gate)
What’s speculative vs. asserted
- Asserted: the six-tool toolkit definition and its “to be built” status (from the 17 Mar 2026 service-line note); the transition-plan scheduling; the market statistics (each carried with its original S-tag).
- Speculative: the “independent evaluation being formalised” market signal remains an interpretive read of converging EU/UK signals, not a formal supervisory requirement [speculative — S-2026-06-19-weekly-briefing]; wave effort estimates (≈1 Claude session per tool) are planning assumptions.
Topics this feeds
- Practice Build — Phase 1 and Phase 2 Tools — turns the “to be built” IGA tool list into a sequenced, briefed build plan.
- Service Line — Independent Governance Assurance — the delivery toolkit this plan builds.
Open questions raised
- Whether the findings register + action tracker should ship to clients as one workbook in every engagement type, or split for larger clients — to be tested in the Wave 3 pilot review.
- Build-completion tracking: the plan’s record-keeping rule is designed to close the known gap that the corpus does not record completion against plan; adherence to be verified at each build session.