You’re asking the wrong question about AI

What should we do with AI?

It sounds like a responsible starting point, doesn’t it? Like you’re taking the latest technology seriously by exploring your options. But in my experience, this question produces a specific kind of response: searching for tools, a pilot or two, or a list of pain points to automate.

This issue is that this framing pulls you toward your current state. You begin thinking about the friction and obstacles that exist today, and look for places to apply AI on top of them. While this is useful work, it’s not enough.

What could be different?

It moves from “how do we fix what’s broken” to “how would this actually work if we designed it for today”? It’s a subtle shift in language, but it creates space for imagination. 

I saw this play out recently with a client. Decades of institutional knowledge were filling multiple systems, from databases to spreadsheets. Answering even a basic question could take hours and meant tracking down the right person. The team didn't open the conversation with "where should we use AI?" They asked: “What would it look like if this knowledge were instantly accessible to everyone who needs it?”

That question led to an AI assistant, yes. They piloted an assistant that lets staff ask questions in plain language and get answers pulled directly from their existing data and documents, all cited and accurate. But they got there because they started with a business question, not a technology one. The AI was the answer, not the starting point.

You can apply that sequence too. Be specific about what you need to be different. Look at what’s actually producing that outcome, then ask where (or if) AI fits.

For a CEO, that might mean reconsidering how value gets delivered to customers, looking at the business model itself. For a functional leader, it might look like asking what their department could be if it were designed from scratch with modern tools and capabilities. The question scales to wherever you sit.

Mid-market organizations don’t have the luxury of long strategy cycles. You know that better than anyone. The goal isn’t to pause everything or stop taking action. It’s to capture near-term value, while also keeping an eye on what a more fundamental change could look like.

A few questions you can think about:

What would the 2027 version of this function or process look like?

Where are we solving for today's constraints instead of tomorrow's opportunity?

If we were building this from scratch, what would we keep?

One more thing: AI is one lever, not the whole answer. Making meaningful change in any part of a business tends to require action across people, processes, data, and technology together. AI can certainly accelerate what’s possible, but the reinvention is broader than the tool.