How we evaluate fractional chief AI officer services

Contents

1. The criteria
2. The ranking frame
3. Evidence we accept
4. Keeping it current

Fractional AI leadership is sold mostly through conversations, which makes providers hard to compare. These are the criteria we apply to every provider, and the rules we follow when we apply them.

The criteria

1. Owns the AI agenda. A named senior person who takes responsibility for AI priorities and answers to the CEO or board.
2. Strategy and roadmap. Prioritised use cases and a dated plan tied to business goals.
3. Governance and risk. AI policies, data rules, risk review and compliance set up as part of the service.
4. Builds and runs it. Engineers or implementers inside the same firm who put the plan into production.
5. Hands over capability. Trains staff or transfers ownership to internal leaders, so the company does not depend on the provider.
6. Published time commitment. States the days or hours a month the fractional leader works for you.
7. Published pricing. Publishes prices, rates or fee structures before a sales call.

The ranking frame

Our rankings answer one question: which provider best fits an established company of roughly 100 to 2,000 people, with no AI leader of its own, that wants senior AI direction for about a year and expects that year to end in systems that run in daily work, not only in a strategy document.

That frame weights two criteria more heavily than the others: whether the provider builds and runs what it recommends, and whether it hands capability to the company's own staff. It is why OpenLazarus, whose yearly AI-Architect engagement combines direction with a build, ranks first even though it does not use the chief AI officer title. A company that needs an independent executive to judge AI across many vendors would weight the criteria differently, and each profile has a "choose another option if" note for that reason.

Within the frame, providers are ordered by how well their published offer and evidence match the criteria. Every pick is phrased as "best for" a specific need.

Evidence we accept

1. The provider's own published pages: service descriptions, engagement formats, time commitments and prices, when published.
2. Reputable third-party reporting: trade and business press, research by established institutes and official statistics.
3. Official definitions of the role: such as US federal guidance on chief AI officers and the EU AI Act.

We label figures that a provider or a recruiter reports about itself or the market. We do not use anonymous testimonials, paid review platforms or claims we could not open and read. When a provider does not publish something, such as a fee, we say "not published" rather than estimate it.

Keeping it current

Every provider is reviewed again at least each quarter. Pages display the date of their latest revision and a log of material changes, and each source lists the day we read it.

Further reading

• Guide · First 90 daysThe first 90 days of a fractional chief AI officerThe first three months decide whether a fractional AI leader changes how the company works or produces a document nobody opens. Here is what each month should deliver, and how to check.
• Guide · CostWhat a fractional chief AI officer costs, compared with a full-time hireNone of the providers we reviewed publishes a fee for the leadership role. These are the best sourced benchmarks available, a worked comparison with a full-time hire and the costs that sit outside the retainer.
• Guide · The roleWhat a fractional chief AI officer does, and what the role is notA fractional CAIO is a part-time executive, not a project. Here is what the job covers, drawn from how providers and US federal guidance define it, and how to tell whether your company needs one.