Why Your IT Department Shouldn't Lead Your AI Strategy

I've spent 30 years inside enterprise technology. ERP implementations at Fortune 500 companies. AI transformation programs. Cybersecurity buildouts. Digital transformation roadmaps across hundreds of client engagements.

And right now, I'm watching a costly pattern repeat itself at companies of every size.

They're handing their AI strategy to IT. And it's costing them.

IT Is Not Built to Answer the Most Important Question

When an organization decides to "do AI," the first instinct is to route it through the technology department. Makes sense on the surface -- technology lives in IT, AI is technology, hand it off.

But here's what happens next.

IT does what IT is built to do. They evaluate security risks. Assess compliance requirements. Build an implementation framework. Run vendor comparisons. Think about infrastructure, data governance, and integration architecture.

All of that work is legitimate. It's necessary. You will absolutely need it.

But not one piece of it answers the question that actually matters:

What do we want to be able to decide or do that we can't do right now?

That's a business question. And when IT is writing the AI roadmap without a business owner in the room asking that question first, you get what most companies have: technically sound AI infrastructure that doesn't move the business forward.

McKinsey has reported that over 80 percent of AI pilots fail to reach production. That's not a technology failure. That's a strategy failure. The technology worked. The wrong problems got solved.

The Structural Fix That Changes Everything

The companies seeing real returns from AI start in a different place.

Before any tool gets evaluated. Before any vendor gets invited in. Before IT writes a single line of architecture. A senior business leader asks: what decisions are we not making today that we should be? What work are we doing manually that's costing us time, accuracy, or competitive position? What data are we sitting on that we've never been able to act on fast enough?

Once that question is answered honestly, everything else follows.

Now IT has something to build toward. Security and compliance serve the business outcome instead of defining it. The roadmap has a destination. The ROI case writes itself because it's anchored to a real problem that real business leaders care about solving.

I've seen this work at 50-person companies and 5,000-person companies. The pattern is consistent. The companies getting ROI from AI share one thing: a senior business leader who owned the "why" before anyone touched the "how."

That leader doesn't need to understand large language models or machine learning pipelines. They need to know the business well enough to answer that one question -- and stay in the room until every line of the roadmap traces back to it.

The Alignment Problem Nobody Wants to Talk About

Here's the harder conversation.

AI doesn't fix organizational dysfunction. It amplifies it.

The same miscommunication that's been slowing down every IT project for the last decade will slow down your AI roadmap. Faster, because the stakes are higher and the spend is bigger.

When I work with companies stalling on AI adoption, the technology is rarely the issue. The issue is that no one in a senior role has built the bridge between what the business needs and what the technology can deliver. Business and IT are running parallel tracks that never quite connect. Pilots get launched without a clear owner. Governance gets bolted on after the fact. Nothing scales.

The fix is a person. Someone who understands enterprise business deeply enough to ask the right questions, and understands modern AI well enough to know what's actually possible. A Chief AI Officer. A Fractional CIO with AI transformation experience. A CIO who already operates at the business level and just needs the mandate to run the program differently.

The title matters less than the function. You need someone who can translate between those two worlds and stay accountable to outcomes, not just deliverables.

The Question Worth Asking This Week

Before your next AI planning meeting, ask your leadership team this:

Does everyone in this room agree on what successful AI adoption looks like for our organization?

Write down what each person says. If you get five different answers, you've found your real problem. And it has nothing to do with the technology.

I challenge you to have that conversation before you buy another tool, fund another pilot, or extend another IT-owned roadmap. Get aligned on the "why" first. Everything else will follow.

To your growth and prosperity.

Mike Martin is the Founder and CEO of Top Down Strategies, Inc. He has led AI transformation, ERP, and digital strategy engagements for over 250 clients across 30+ years. He is a TEDx speaker, published author, and Fractional CIO/CTO serving mid-market and enterprise organizations.

Ready to align your AI strategy to business outcomes? Book a conversation with Mike.