The CIO’s AI dilemma: What to do when the Board says “Use AI”

If you’ve sat in a board meeting recently, you’ve probably heard some version of this.

“What’s our AI strategy?”

Or perhaps:

“Our competitors are investing in AI. What are we doing?”

I’ve been on the receiving end of those conversations myself. As CIO at Stagecoach, I knew AI could make a real difference to the business. I also knew there was no shortage of consultants, vendors and opinions about where to start. After plenty of workshops and presentations, the responsibility still came back to one person: deciding where AI would deliver value, where it wouldn’t, and how to move forward without wasting time or budget.

That’s the position many CIOs find themselves in today.

This article shares the approach I’d take during the first 90 days after the board asks for an AI strategy, drawing on the lessons I learned implementing AI inside a large organisation and the conversations I’m now having with leadership teams across different sectors.

In this article, you’ll learn:
● Why board pressure around AI leaves many CIOs caught between delivering results and managing risk.
● How to approach your first 90 days with a clear plan instead of jumping straight into technology.
● Which early AI initiatives are most likely to deliver measurable business value.
● How to build governance and manage risk without slowing progress.
● How to report AI progress to the board with confidence and credibility.

What the board is really asking
When a board asks for an AI strategy, they’re usually looking for reassurance as much as innovation. They want to know the organisation isn’t falling behind competitors. They want confidence that opportunities aren’t being missed. And they want to understand the risks, the investment required and what success could look like.

I’ve seen organisations rush into vendor demonstrations because the board wanted momentum. Others launched pilots across multiple departments at once, only to discover six months later that nobody had agreed how success would be measured or who would own the projects beyond the trial.

It’s easy to see how that happens. AI vendors all promise transformational results. Every department has ideas. Everyone wants to be first with the next breakthrough.

But a CIO has to take a different view.

Before looking at products, I always want to understand where the business is trying to improve. Which processes consume the most time? Where are costs increasing? Which customer frustrations keep appearing? Which teams are overloaded with repetitive work? Those conversations produce a much stronger shortlist than any technology demo.

That’s also the point where AI becomes part of a wider business conversation instead of a standalone technology initiative.



Research suggests many organisations are still working through this challenge. The Kyndryl Readiness Report 2025 found that 74% of CEOs say they are not aligned with their CFO on long-term value of technology investments. And 70% of CEOs report they got to their current cloud environment by accident, rather than by design. 

Those numbers don’t surprise me. Deciding which ideas deserve investment, bringing people with you and creating enough governance to support long-term adoption takes much more thought.

From there, the conversation changes. Instead of trying to answer, “How can we use AI?” you’re looking at a much shorter list of business problems where AI has a genuine opportunity to make a measurable difference.

The first 30 days: build a clear picture before making big decisions
One of the biggest surprises for many leadership teams is how much AI activity is already happening before an official programme even exists.

Someone in marketing has been using ChatGPT to draft content. Finance have started experimenting with Copilot. Developers are testing new coding tools. Customer service have looked at AI assistants. It all starts with good intentions, but very few organisations have a complete picture of what’s being used, why it’s being used, or whether anyone has considered the risks.

That’s why I’d want to know which processes are causing the biggest headaches. Where people spend hours every week on repetitive work. Which teams are struggling with backlogs. Where customer complaints keep appearing. Those conversations tell me far more than a list of AI products ever could.

At the same time, I’m looking for AI that’s already found its way into the organisation.
Sometimes it’s completely harmless. Sometimes it creates security, compliance or governance concerns that nobody realised existed. Either way, it’s better to know about it early than discover it halfway through an implementation programme.