AI maturity model: five levels and how to move up
An AI maturity model shows how far a company has moved from experiments to controlled use in core work. Paul Okhrem uses five levels and six dimensions. Place your company on the model, then test that view with the free AI readiness scorecard.
The five levels of AI maturity
The five AI maturity levels are exploring, experimenting, operational, scaled and embedded. Judge each level by evidence, not the number of tools purchased.
| Level | Name | What you see |
|---|---|---|
| 1 | Exploring | People use tools individually. There is no shared AI inventory, policy or owner. |
| 2 | Experimenting | Several pilots run. A first acceptable-use policy exists. Reported results are mainly anecdotes. |
| 3 | Operational | Some systems run in production. Named owners, risk tiers and baselines cover priority use cases. |
| 4 | Scaled | A funded portfolio passes release gates. Teams share evaluation standards, controls and regular board reporting. |
| 5 | Embedded | AI supports core processes under clear authority. Teams monitor quality, cost, risk and portfolio value. |
A company may reach different levels in different workflows. A controlled low-risk use case does not establish readiness for a higher-risk system.
Six dimensions to assess
Assess strategy, data, governance, delivery, adoption and value measurement separately. A strong technical platform cannot replace business ownership or a clear evidence trail.
| Dimension | Level 1 | Level 3 | Level 5 |
|---|---|---|---|
| Strategy | No shared priorities | Ranked use cases with owners | AI choices form part of business strategy and funding |
| Data | Ad hoc access | Priority data is accessible and quality is known | Owned data products support multiple systems |
| Governance and risk | No policy | Inventory, risk tiers and release reviews | Ongoing evidence and controls across the lifecycle |
| Delivery and platforms | Individuals choose tools | Standard platforms support priority cases | Shared evaluation, monitoring and support |
| Adoption and skills | Individual curiosity | Role-based training for named workflows | AI use is reflected in job design and operating goals |
| Value measurement | Anecdotes | Baselines for key use cases | Portfolio costs and measured value tracked against plan |
How to move up one level
Close the next ownership or evidence gap before buying more tools. Use one bounded workflow to test the change.
About 6% of organizations qualify as AI high performers, attributing 5% or more of EBIT to AI. McKinsey's November 2025 report also required respondents to report significant value from AI. EBIT means earnings before interest and taxes. The self-reported survey covered 1,993 participants in 105 nations. This definition is separate from the five planning levels on this page.
- Exploring to experimenting: choose a small use-case shortlist, publish an acceptable-use policy and name an executive sponsor.
- Experimenting to operational: build the inventory, assign risk tiers and record a baseline. Release one justified use case after testing.
- Operational to scaled: fund a portfolio, standardize evaluations and monitor results. Give the board a regular decision report.
- Scaled to embedded: redesign suitable processes, assign operating owners and track value across the portfolio.
Use an AI policy template for permitted use and an AI transformation roadmap for phased decisions. Neither replaces the evidence required for a particular release.
AI maturity model versus AI governance maturity model
This model covers the whole AI capability. The AI governance maturity model focuses on controls and decision rights.
Use both when ownership, risk review or evidence is the main gap. Maturity is a planning tool, not a compliance certificate.
Assess your level
Start with the free 28-question AI readiness scorecard. Record evidence for each answer and name the owner of each gap.
A fractional Chief AI Officer can lead an agreed improvement mandate when executive ownership is missing. A fractional chief data officer addresses data ownership and readiness. AI implementation consulting helps turn an approved use case into controlled daily work. Paul Okhrem’s services are separate from this free self-assessment.
This is Paul's practical framework, not a ranking of companies or a validated prediction of financial returns. Keep a review date and revise the assessment when evidence changes.
FAQ
What is an AI maturity model?
An AI maturity model places a company on a scale from experiments to controlled use in core processes. It identifies the next ownership, data and operating gaps to address.
How many AI maturity levels does this model use?
Five: exploring, experimenting, operational, scaled and embedded. Assess the six dimensions separately before assigning an overall level.
How long does it take to move up one level?
There is no fixed timetable. The work depends on the starting point, risk and scope. Ownership, data and process changes may take longer than installing a tool.
Who should own AI maturity?
One executive should own the overall plan. This can be a Chief AI Officer, another executive with a written AI mandate, or a fractional CAIO. Workflow owners retain their operating responsibilities.
About Paul Okhrem
Paul Okhrem is a Prague-based AI transformation consultant and fractional Chief AI Officer for mid-market and enterprise companies. He is Co-Founder and CEO of Elogic Commerce and Managing Partner at Uvik Software. He has built B2B and enterprise software since 2009.
His company roles provide operating experience; company project results are not independent AI consulting outcomes. Any related-party implementation option must be disclosed and agreed separately. Read the evidence and measurement method.
Change log
- : Reviewed copy, source boundaries and supporting resources.
- : Reviewed service scope, source evidence, links and buyer answers for the October release.
Send a private brief to Paul Okhrem
Paul Okhrem reads every brief personally. Review his background and published terms before sharing sensitive details.