Issue #15 June 14, 2026

Codex vs. Claude Code, What AI Is (and Isn't) Learning From You, AI Adoption Strategy at Its Core — and How Legal, Risk & Compliance Are Thinking About AI Transformation

Editor's Take

Three AI stories this week: why Codex deserves a second look — especially if you're not a developer; whether your AI actually learns from you (the model doesn't, the notebook around it does); and a brief on the U.S. government forcing Claude Fable 5 & Mythos 5 offline. Then two Org stories: a distilled AI-adoption strategy, and an interactive briefing on AI for Enterprise Legal, Risk & Compliance.

This Week's Briefings

Two Org stories this week: a tight, practical AI-adoption strategy, then a deep interactive briefing on AI for Enterprise Legal, Risk & Compliance.

Disclaimer: All opinions in these briefings are my own and do not represent any company's position. This is meant as an academic discussion. Firm-reported figures are company or vendor commentary, not independently audited fact.

1 · AI Education & Adoption Strategy: Three Insights at the Core

There are countless AI-adoption strategies and playbooks out there. At its core, though, an effective AI-adoption strategy — and the insights behind it — distills to just three points: remove the basic-skills blocker, harness peer-to-peer learning, and make wins visible, relevant, and bite-sized. These build the foundation for everything that follows in AI transformation — especially workflow redesign and operating-model reimagination.

#InsightStrategySuccess looks like
1 The basics are the blocker. Many people don't know how to even find Copilot — and not knowing how to prompt stops them from using it at all. Cover the ground with bought training. AI literacy firm-wide, Copilot training, executive education, and vendor-delivered functional training — credible external content, deployed fast, so the basics are never the reason someone isn't using AI. Everyone is in the game. Foundational literacy completed firm-wide; active users grow from X to Y (a% to b% of licensed users).
2 Peer-to-peer is the most effective adoption mechanism. Microsoft's own research found most employees skip formal onboarding material and learn AI socially, from colleagues — not from courses or mandates. Identify, build, and grow AI Frontiers & Champions. Give them deeper AI knowledge and tools, self-paced learning, power gather-togethers (AI Frontiers Weekly), and the role mandate and time to spread it. Usage deepens, not just spreads. Champion coverage across every major function; weekly-active intensity rises from X to Y users active every week of the month.
3 Adoption compounds when wins are visible, relevant, and bite-sized. A 1-minute peer demo in your own role's workflow converts more skeptics than any mandate or hour-long course. Viva Engage as the amplification platform — and the home of built training. 1-minute videos, real role-based use-case sharing, and function-specific training built by AI Frontiers in each function's own workflows and language. The flywheel turns. Copilot-assisted hours grow from X to Y per month; a growing set of use cases logged, the best progressing into embedded practice.
The throughline: literacy removes the excuse, champions create the social proof, and a low-friction amplification channel turns isolated wins into a flywheel. None of the three works alone — bought training without peer amplification stalls at "everyone took the course, nobody uses it," and champions without a platform stay invisible.

2 · AI for Enterprise Legal, Risk & Compliance: The Dual Mandate

An executive briefing for the GC, CRO, CCO, CPO, Chief Audit Executive, CISO, and the board.

Heads-up: The interactive briefing dashboard embedded below is the main deliverable — nine argument sections plus twelve expandable reference tables.

Legal, Risk, and Compliance carry a tension no other function does: they both adopt AI and must police everyone else's use of it. The functions pulling ahead treat that dual mandate as the design principle — the same discipline that makes them effective adopters (inventory, validation, audit trails, human-in-the-loop) is exactly what the enterprise needs from them as AI's control authority. The briefing below is the full argument and evidence layer.

The dashboard — The Dual Mandate: Using AI & Governing It — is the full evidence layer: the thesis, the adoption-vs-understanding paradox, four levels of workflow redesign, enterprise cases, a five-level maturity model, the four-tier AI red-lines framework, the 2024–2026 regulatory timeline, and a 12–24 month plan — with twelve expandable reference tables underneath (including one with no analog in a finance dashboard: governing enterprise AI). Open it in a new tab if it loads slowly.

Stop measuring AI maturity by how many tools the function has adopted. Start measuring it by the question that defines the dual mandate: can you produce, on demand, a complete inventory of the enterprise's AI — including the generative and agentic systems your regulator just placed outside the model-risk perimeter — and the evidence that each is governed? By 2027, every function will have the same frontier tools. The difference will be defensibility. Using AI well and policing it well are not two jobs. They are one.

Get the newsletter in your inbox

A weekly, curated digest on AI and organizational design.