Fractional chief AI officer: what it is and when you need one

What a fractional chief AI officer actually does

Strip out the title and the job is decision-making. Someone senior has to say which of the fifteen possible AI projects your business could run is the one you run this quarter, and then defend that choice when a shinier one turns up six weeks later. In a large company that person is on payroll. In a business of five to fifty people the work is real but it is not a full-time role, so you rent the judgment instead of hiring it.

• Sets the order: which problem gets attention now, which waits, and which gets dropped for good.
• Picks the tools and says no to the rest, including the ones the team is excited about.
• Writes the rules: what staff may put into an AI tool, what stays out, and who signs off on anything that reaches a client.
• Briefs and evaluates vendors, so you are not comparing three quotes written in language none of you can read.
• Owns the honest measurement: did the last change save time, and should you keep paying for it.
• Keeps you off the projects that demo beautifully and return nothing.

What the role is not

It is not a builder. Advice is the product. A fractional chief AI officer who also sells you the build has an obvious reason to find one, and you lose the only thing you were paying for. In my case, if a build turns out to be the right answer, that is a separate engagement with its own scope, price and contract, and you are free to take the plan to anyone you like.

It is also not a governance exercise. If what you need is a policy document and a committee, a consulting firm will produce one. This role is closer to having a technical co-founder for fortnightly consulting calls with written next steps.

When ongoing advice makes sense

The test is how often the decisions arrive. If you have one question to settle, pay once and settle it. If a new question lands every fortnight because you are actively rolling things out, hiring, or being quoted by agencies, a retainer costs less than the mistakes it prevents.

What it costs

AI Consulting is A$1,195 per month + GST, with a call each fortnight, clarity on what to build in Claude or ChatGPT and practical guidance on how to do it. No lock-in.

Custom development is A$150/hour + GST, with a separate project scope, estimated hours and terms before work starts.

We agree the work, estimated hours and engagement terms before starting.

How to evaluate one

The market for this title is about eighteen months old, which means the average level of proof behind it is low. Six questions sort it out quickly.

1. 1Do you sell, resell, or take a commission on any of the software you recommend?
2. 2What have you personally built or shipped in the last six months? Ask for specifics, not a client logo wall.
3. 3What would you tell me not to do? A good answer arrives fast and costs them money.
4. 4What happens between calls, and who owns the actions when we hang up?
5. 5How do I stop, and how much notice do you need?
6. 6Can I see what one of your written action lists looks like?

What a fortnight actually looks like

Bring what is actually in front of you: a quote from an agency, a tool you are considering or a process that is falling over. We work through the options and agree the next step. The scope and estimated hours are agreed first.

Between calls, you work through the agreed next steps. Bring what happened, what got stuck and any new questions to our next conversation.

If what you want is closer to a standing sounding board than a chief anything, the AI business coach framing describes the same arrangement in plainer language. If the pressing problem is that a rollout has stalled, start with AI adoption consulting instead.