Field-tested by a practicing AI Champion

You got the AI mandate.
Nobody handed you a manual.

The strategy books talk to your CEO. The prompt books talk to your users. This is the operator's manual for the person in between, the one who actually has to make AI adoption happen.

Second edition, August 2026. Errata and the quarterly vendor annex: kesslernity.com/playbook

Get the Playbook for $59 See what's inside
139 pages, field-tested 12 working artifacts EPUB, PDF, and editable Word
88%
of organizations now use AI in at least one function
~39%
report any enterprise EBIT impact from it
~95%
of enterprise gen-AI pilots show no measurable P&L impact
27 pts
leaders vs employees on basic AI-agent familiarity (67% vs 40%)

Sources: McKinsey State of AI (88% / ~39%); the MIT NANDA report, a lab working paper rather than a peer-reviewed study, on the ~95% with no measurable P&L impact. Adoption went wide. It did not go deep. The value leaks at the operator layer, which is the seat this book equips.

The seat nobody writes for

You have more competence than authority, and more impatience than time.

You sit between a slide that says "we are doing AI" and a team already pasting work into a chat window. You own the gap, and there was no manual for it. So here is the work, in the order you will actually face it.

The pilot that proves nothing

You ran a pilot, but it produced a demo, not a number. You cannot defend "people seem to like it" to a CFO at renewal.

The gate that stalls you

Legal and security can stop you with one email, and you do not know what a DPIA wants or where the lawyer takes over.

The number that holds up

Licenses assigned is not adoption. Logins is not adoption. You need a unit of value that survives the room. This book builds it.

What you get

A read that becomes a toolkit.

The book teaches the thinking. The template pack does the doing. You walk into your next sponsor meeting with the artifact already drafted.

The book

139 pages, 13 chapters across 5 parts

Second edition, August 2026. Delivered as EPUB and PDF, so it reads on your e-reader, your phone, or your second monitor. Every chapter ends with a move you can make on Monday.

The work in order

Scope a use case, score it, charter the pilot, clear the gate, train the team, prove the number, scale what works.

6 appendices you will pull from repeatedly

A template manifest that maps "I am on step X" to "pull template Y"; a five-axis maturity self-assessment; a glossary that keeps the room speaking the same language; a jurisdiction map covering the EU, UK, US states, the UAE and Gulf, and Singapore; the full research bibliography, printed, so every citation in the book resolves; and a curated reading list.

Sourced and dated, honestly

Every stat carries a citation. Field experience is labeled "field." Where the data has no clean number, it says so instead of inventing one.

The 12-artifact pack

Working artifacts, not reading material

You do not retype the frameworks, you fill them in. Shipped two ways: editable Microsoft Word (.docx) with real tables and checkboxes you complete directly, and the markdown source to drop into Notion, Confluence, or any tool in your stack.

The three kinds of notes

The book flags the moves the way you will use them:

Field Note Trap Move
The template pack

Eleven templates, each tied to the chapter that teaches it.

Read the chapter, open the template, fill it in. Appendix A maps every one to its anchor chapter and what it produces.

1Executive Sponsor One-Pager
Chapter 2
2Use-Case Intake & Scoring Sheet
Chapter 3
3Pilot Charter
Chapter 4
4RACI
Chapter 4
5Training Curriculum Outline
Chapter 5
6Adoption KPI & Dashboard Spec
Chapter 8
7DPIA & Security Pre-Flight Checklist
Chapter 6
8Communications & Newsletter Cadence
Chapter 9
9AI Usage Policy Starter
Chapter 9
10Maturity Self-Assessment
Chapter 11
11Tool Selection Rubric (8 dimensions)
Chapter 7
+All editable. All yours to adapt.
Word + markdown, inside your org
Fit check

Built for the operator. Not for the deck.

This is for you if

  • You got the AI mandate and need a path through the work.
  • You lead AI enablement and are building a program from a blank page.
  • You are an IT or innovation lead who owns the pilot, the security review, and the rollout.
  • You are the accidental AI lead handed the title last quarter, who would like a manual now.

Not for you if

  • You are a C-suite exec who wants a vision deck. Your strategy book already exists.
  • You want a list of clever prompts. Buy a prompt book; this operates one level up.
  • You want theory without the work. Every chapter asks you to do something.
The credibility is not borrowed
Mathieu Kessler

Written by a practicing internal AI Champion, not a consultant. Mathieu ran a real enterprise pilot on Claude Enterprise, mentored leadership through their own GenAI adoption, personally cleared the DPIA and the security review, and publishes an internal program newsletter, failures included. He carries a FinOps background, so the cost and impact chapters read like someone who has defended a number, because he has. Field-tested, not theorized.

Questions

The honest answers

What format is it?
EPUB and PDF, both included in the download. The pack's 12 operational artifacts (11 editable templates plus the tool-selection scoring rubric) ship in two formats: editable Microsoft Word files (.docx) you fill in directly, and the plain-text markdown source you can drop into Notion, Confluence, or any tool in your stack.
The EU AI Act keeps moving. Will the book go stale?
It already did once, and that is why this edition is built differently.

The first edition printed the enacted 2 August 2026 high-risk date with a warning that the Digital Omnibus would probably defer it. It did. Regulation (EU) 2026/1744 entered into force on 27 July 2026, moving the standalone high-risk regime to 2 December 2027 and embedded systems to 2 August 2028, while leaving the Article 50 transparency duties live from 2 August 2026. I corrected the chapter and pushed the update free to everyone who had already bought it.

The second edition separates what perishes from what does not. The method, the rubrics, the failure modes and the KPI design are in the book and hold for years. The fast-moving vendor facts behind the Chapter 7 rubric live in a versioned annex at kesslernity.com/playbook, refreshed quarterly, alongside the errata and change log. When a date or a vendor fact moves materially, I correct the digital files and notify existing buyers.
Does it cover AI agents and MCP servers, or just chat assistants?
Chapter 8 is dedicated to it. Connectors, MCP servers and browser agents are the requests landing on champions right now, and the vendor paper trail you rely on for an assistant does not exist for a community-maintained connector. The chapter gives you the review that replaces it, five checks: provenance, scope, hosting, permission inheritance, and the sign-off routing. Plus the credentials and outbound-flow rows that go on the same sheet, and how to handle agents that carry their own standing credentials instead of inheriting the user's. Written by someone who runs MCP servers in production.
What license do I get?
A single-user license for personal and internal professional use. You may use and adapt the templates inside your own organization. You may not resell, redistribute, or republish the book or the pack.
Can I buy it for my whole team or site?
Yes. For team or site licensing across multiple seats, reach the author through linkedin.com/in/mathieukessler and we will sort a volume arrangement that fits your headcount.
Get the playbook

Open the sponsor one-pager template, and draft it before your next meeting.

  • The 139-page book, EPUB and PDF
  • All 12 artifacts (11 templates plus the scoring rubric) in editable Word and markdown
  • Free updates as the rules change
  • Instant download, yours to keep
$59
One-time payment. Instant download of the book and all 12 artifacts.
Buy now and start Monday

Stay curious, stay clever.