Everyone In Your Business Uses AI. Nobody Owns It.

Ask a leadership team who is responsible for AI in their business, then watch the room.

The IT manager looks at the COO. The COO looks at the CEO. Someone mentions a person in marketing who is “really into it”. Then somebody says the word committee, and everyone relaxes.

I’ve seen some version of that in a lot of leadership rooms. Meanwhile the team is using AI every day. Drafting emails, summarising contracts, writing reports, analysing spreadsheets that might hold client data. Leadership doesn’t know where it’s being used, on what data, or what it’s deciding. Nobody is being careless. There’s just no one whose job it is to know.

Why It Falls Between Chairs

AI doesn’t fit the org chart. It’s software, so it looks like IT’s problem. It changes how work gets done, so it looks like operations. It needs rules, so HR and legal get pulled in. It costs money, so finance has a view.

When something belongs to four departments, in practice it belongs to none of them. Each one does its slice. IT approves a tool. HR circulates a policy. Finance queries the subscriptions. Nobody looks at the whole picture or asks whether any of it is making the business better.

Committees feel like the answer, and in my view they rarely are. A committee meets monthly, produces minutes and waits for someone else to act. AI is moving faster than a monthly meeting cycle. You need one person who wakes up thinking about it.

IT Is The Wrong Default

This one surprises people, so let me be careful. IT absolutely has a role. Security, access, approved tools, data controls, the integrations. That’s critical work and it belongs with them.

The value of AI lives somewhere else. It lives in the workflows. The tender that takes a week. The site report written up at the office at 7pm. Month-end. The client brief that gets rewritten three times. Whoever owns AI has to understand that work closely enough to see where the hours are going, and has the standing to change how it’s done.

Hand the whole thing to IT and you tend to get well-secured tools that nobody uses well. Hand it to an operations-minded leader working hand in glove with IT and you get both.

What The Job Actually Involves

Give the role a short, concrete job description. Five things.

Keep the register. A living list of where AI is used across the business. Which tool, for which task, on what data, and who reviews the output before it matters.

Pick the first workflows. Two or three, chosen by hours saved and risk reduced. Ignore the shiny ideas.

Set the rules. With IT on the tools and a lawyer on the wording. The owner makes sure the rules match how people really work, which I go into in writing an AI policy people follow.

Get people trained on their real work. Generic prompt courses don’t stick. Training on your own tenders, reports and client tasks does. There’s more on that in building AI capability.

Report every month. To the leadership team, on one page. Hours back. Issues found. What’s next. Treat it like any other investment and measure it that way, as I set out in how to measure AI ROI.

Who Makes A Good Owner

Look for three qualities, and seniority on its own isn’t one of them.

Curiosity first. The right person has already been playing with these tools on their own time and has opinions about them. Second, enough authority to change how a team works without asking permission for every step. Third, the trust of the front line. If the people doing the work see the AI owner as head office checking up on them, they’ll hide what they use and you’re back where you started.

Often that’s an operations leader. Sometimes it’s a general manager or a senior project lead. Rarely is it the most junior enthusiast, however keen they are, because they can’t move a department.

Then give them actual time. If AI is squeezed into the gaps of an already full job, it won’t happen. Carve out a fixed part of their week and take something else off their plate to make room. Most businesses of twenty to a few hundred people don’t need a Chief AI Officer. They need a senior person with this written into their role.

The First Month Is The Register

Start by finding out what’s already happening. Ask the team, anonymously if that helps, which AI tools they use and what for. Make it clear nobody gets in trouble for the answer. You want the truth more than you want compliance, and the honest picture is usually bigger than leadership expected. It’s the same ground I cover in shadow AI.

Map each use against the data it touches. Client information. Financials. Staff records. That map tells you where to tighten things straight away and where you can relax.

Then document the workflows that matter, so they live in the business instead of one person’s browser history. In one mining and resources business we worked with, a key person left and nothing broke. That’s the standard to hold your AI owner to. Every important AI workflow should survive the person who built it going on leave.

Put A Name Next To It

At your next leadership meeting, put AI on the agenda for ten minutes. Leave with one name, a set amount of time in that person’s week, and a date for their first one-page report.

If the meeting ends without a name, you’ve still made a decision. You’ve decided that whoever happens to be most enthusiastic will set your AI rules, one browser tab at a time.

Frequently Asked Questions