Written for CIOs, COOs, chief data and transformation officers, and the sponsors who have to decide what to do next.
What's in this framework
This 28-page framework is a workshop diagnostic that uses the same language as our online AI maturity assessment. A readiness score is only useful if it is based on evidence, if two people in the same organization would reach a similar number, and if it tells you what to do next. This framework is designed to do all three.
The model is eight dimensions — Data Foundation, AI Adoption, Talent & Skills, Infrastructure, Governance & Ethics, Strategic Alignment, Process Automation, and Investment — and five levels: Nascent, Emerging, Developing, Advanced and Leading. The overall score is the sum of eight dimension scores, each from 1 to 4, on an 8–32 scale. Culture is not a ninth dimension. It shows up in how the eight are evidenced.
It is written for CIOs, COOs, chief data and transformation officers, and the sponsors who have to decide what to do next. The illustrative mid-size insurer in the document is Emerging, at 15 of 32, with a target of Developing, at about 24, by raising the foundations the pilots will need. It is labeled illustrative. It is not a client case study, and it is not a sector average. A companion workbook holds the 32 indicators, the levels, and the illustrative profile. A completed workbook is a scored position, not a certificate.
How the workshop is meant to be run
Evidence, not opinion. Each indicator is awarded 1 to 4 with a one-line reason and a piece of evidence a second person could examine. A room that scores from slogans has not finished the workshop. Send the eight dimensions in advance. Ask each attendee to bring evidence, not a preference.
The target is the next level, not Leading everywhere. Do not wait to be Leading before you fund a first case. A first case should raise the constraint — often data, governance, talent, or measurement — rather than skip it. The governance framework, the data-strategy guide and the 2026 diagnostic in this library are companions. The online assessment at /ai-maturity uses the same model; the workshop behind the score is what makes the number one you can stand behind.
This framework will not tell you that you are “ahead of your sector.” It will not convert a 15 into a budget. It will not excuse a missing baseline. It will tell you where you stand, on evidence, and what the next honest move is. The 2026 diagnostic is the same model, used as a diagnostic view with illustrative profiles and a sector lens — not a survey.
Eight dimensions, five levels
The same model as our online AI maturity assessment. Each dimension is scored 1 to 4. The overall score is the sum, 8 to 32. The level is read from that sum.
Data Foundation
Named sources, shared definitions, quality that can carry the work, and permissions. The data-strategy guide in this library is the companion.
AI Adoption
What is actually running, on an inventory, with people who use it. Assistants that are already in use are not “informal” and therefore off the score.
Talent & Skills
The people who will design, operate and challenge the work, including an operator who is not only the builder.
Infrastructure
Where the work will run, and whether a workload can leave. The offline guide is the companion when data should not leave.
Governance & Ethics
Inventory, proportionate path, counsel asked in time, a person who can stand behind a decision. The governance framework is the companion.
Strategic Alignment
A named job, a sponsor, and a first case that raises a constraint rather than a slogan.
Process Automation
Processes that are understood well enough to change, and a mechanism that fits the job. The workflow playbook is the companion.
Investment
A signed baseline on a named process, a cost that is complete, and a way to question the number. The ROI guide in this library is the method; this framework will not stand in for it.
The five levels are Nascent, Emerging, Developing, Advanced and Leading. Target the next level. The illustrative insurer is Emerging at 15, with a target of Developing at about 24 — not Leading on every dimension.
How a score is meant to be used
A score that two people in the same organization would not reach similarly is not yet a score. The workshop is the method.
Evidence, not opinion
Each indicator has a one-line reason and a piece of evidence. Award 1 to 4. A missing reason is a reason to pause, not a 3.
Target the next level
The honest move is usually one band, on the dimensions that will make a first case survivable. Do not wait to be Leading.
A first case that raises the constraint
Fund work that raises data, governance, talent or measurement. A first case that skips the low scores “to show value” has skipped the design.
Get the guide and workbook
Enter your name, work email and company. The files appear on this page immediately. A backup download link is also emailed when email delivery is available.