Who Owns Your AI Tranformation? The Question That Predicts Who Wins.

A statement I often make in boardrooms tends to get an immediate reaction:

If your AI team reports to IT, you're already behind.

Not because IT isn't important. Some of the best technology leaders I've worked with have been instrumental in driving transformation. The problem is something else entirely.

Most companies still think of AI transformation as a technology initiative. The companies actually winning with it treat it as a business transformation initiative. That distinction sounds small. It isn't. It's why so many organizations are spending millions on AI pilots and seeing almost nothing move — not revenue, not customer experience, not productivity, not competitive advantage.

How Most AI Transformation Journeys Usually Go

I've watched versions of this story play out inside enough large organizations that I could write the script with my eyes closed.

The CEO hears about AI. The board asks for a strategy. The CIO/CTO is told to look into it. A central AI team gets stood up. A few pilots launch. A chatbot appears on the website. A handful of reports get automated.

A year later, leadership is quietly wondering why none of it moved the needle.

The issue was never the technology. It was ownership. When AI sits entirely inside IT, it tends to become a support function instead of a growth function — the team's attention drifts toward infrastructure, governance, model deployment, vendor selection, security reviews. All of that matters. None of it creates business value on its own.

What Foodpanda Taught Me About This, Before It Had a Name

Long before ChatGPT made AI a boardroom buzzword, we were already using machine learning to improve business outcomes at Foodpanda.

Back in 2013, one of our biggest problems was simple to state and hard to solve: helping a customer find the right restaurant, fast. Most food delivery platforms sorted on distance, ratings, and sponsored listings, and left it at that. We tried something different.

Our product, growth, and data teams built personalized restaurant ranking models together — not as a side project handed to engineers, but as something the business team owned directly. The models looked at order history, cuisine preference, time of day, basket size, local delivery performance.

The goal wasn't to build an AI model. The goal was to increase conversion.

Because ownership sat with the business, every model was judged the way a business decision should be — against click-through rates, conversion, repeat orders, customer lifetime value. The algorithm wasn't successful because it was clever. It was successful because people ordered more food. That's the whole difference, and it's a difference I've carried into every role since.

The Question That Actually Matters

It isn't "who built it." It's who owns the outcome.

Picture two companies. In the first, the AI team reports into IT, and their goals look like pilots launched, models deployed, infrastructure uptime. In the second, the AI team sits inside the business, and their goals look like revenue growth, retention, cost reduction, customer satisfaction.

I don't think there's much mystery in which one compounds. Yet most organizations are still built like the first

Where I've Seen This Work

AI creates value when it improves a decision. Decisions rarely happen inside IT. They happen inside product teams, marketing teams, editorial teams, sales, operations, finance.

At Times Group, we have been working on GPT (Generative Pre-trained Transformer) and BERT (Bidirectional Encoder Representations from Transformers) LLM models before OpenAI launched their chat interface to access GPT which has now become a household tool.

For us, our most successful AI initiatives were never really technology projects. They were business projects that happened to use AI. Editorial teams used it to improve content workflows. Product teams used it to sharpen personalization. Audience teams used it to lift engagement and retention. The technology enabled the work. The business owned what came out of it. I've found that's a distinction worth insisting on, even when it feels pedantic in the room.

Amazon is a useful example, because people assume it has some enormous centralized AI department quietly running the company. It doesn't. AI is woven into pricing, recommendations, logistics, forecasting, advertising — not as a research function sitting apart from the business, but as the business itself. The recommendation engine isn't an IT project. It's a revenue engine. The forecasting system isn't a technology initiative. It's an operational advantage. AI sits where the decisions happen, and that's exactly why it keeps compounding.

Spotify tells a similar story. Its recommendation systems shape billions of listening decisions a day, but the people responsible for them aren't isolated researchers working in a vacuum — they sit with product managers, designers, growth teams, and the people who actually understand the content. Everyone owns the outcome together. The technology exists to serve the business objective, not the other way around.

The Structure I'd Actually Build

If I were designing an AI-first organization today, I wouldn't build one large centralized AI department. I'd build a hub-and-spoke model instead.

The hub is a central platform team, responsible for infrastructure, governance, security, model frameworks, shared capability. The spokes are embedded AI specialists sitting inside product, marketing, operations, finance, customer experience — reporting into business leaders, while drawing on what the hub provides.

That combination gives you both speed and consistency, and I think it's the only structure that actually survives contact with a real P&L.

Take a retail company trying to improve retention. In the traditional structure, the AI team builds a churn model, the model gets deployed, a report gets generated, and — more often than anyone likes to admit — nobody uses it. In the structure I'm describing, the AI specialist sits with the CRM team from day one. The churn model flags at-risk customers, marketing launches personalized campaigns automatically, support reaches out proactively, product teams dig into why people are actually leaving. The model becomes part of how the team works, not a report they occasionally open. Same technology. Completely different outcome.

What This Means for IT

None of this makes IT less important. If anything, it raises the stakes on the role.

The future CIO becomes one of the most strategic seats in the company — not because they own every AI initiative, but because they enable AI across all of them. They provide the platforms, the governance, the security, the data infrastructure. The business provides the priorities, the use cases, the outcomes, the accountability. Transformation happens where those two things meet, not where one substitutes for the other.

Where I've Landed

The biggest misconception about AI right now is that it's a technology problem. It isn't. It's a leadership problem, and underneath that, it's an organizational design problem.

The companies building a lasting edge with AI aren't necessarily building better models. They're building better operating systems around the models they have.

So when someone tells me their company is AI-first, I ask one question: who does your AI team report to?

The companies winning with AI aren't building better models. They're building better operating systems.

The answer usually tells me everything else I need to know.