Your Platform Readiness Score

Insights

AI Isn’t Your Biggest Problem. Your Data Is.

At the Retail Technology Show 2026, Nimitt Desai, Head of Innovation, AI & Technology at PMC, cut through the noise around AI with a message that landed hard:

AI isn’t the problem. Data is.

After 25 years working across major technology shifts, from monolith to microservices, cloud to serverless, Nimitt has seen what successful transformation looks like. But what’s happening with AI is different.

“This isn’t just another technology shift,” he explained. “It’s an ecosystem transformation, and most organisations are underestimating what it really takes to make it work.”

The illusion of AI readiness

Many retailers believe they’re ready for AI.

They’ve invested in modern infrastructure. They’ve built capable teams. They’ve identified high-value use cases.

On paper, everything looks right.

But in practice, things fall apart, not because of the AI itself, but because of what sits underneath it.

“The biggest problem isn’t the technology,” Nimitt said. “It’s the data foundations, and they’re often the most ignored part of the journey.”

When AI goes live, and goes wrong

To bring this to life, Nimitt shared a familiar retail scenario.

A well-established retailer, strong in both domain expertise and technology, decides to launch an AI-powered product recommendation engine. The use case is solid. The proof of concept performs well. Confidence is high.

So they go live.

And almost immediately, the experience breaks.

Customers are shown irrelevant products. They’re recommended items they’ve already purchased. They’re encouraged to buy products that are out of stock.

Instead of driving revenue, the experience frustrates users.

The assumption? The AI has failed.

The reality? The data has.

“The AI was doing exactly what it was designed to do,” Nimitt explained. “But it was working with fragmented, inconsistent, and outdated data.”

AI doesn’t fix bad data. It amplifies it.

This is where many organisations get it wrong.

AI is often seen as a solution layer, something that can clean up, optimise and improve what already exists.

But that’s not how it works.

“AI cannot fix fundamentally broken data,” Nimitt said. “It amplifies what exists. If you have clarity, it scales it. If you have chaos, it exposes it.”

In many retail environments, that chaos comes from a familiar place:

Multiple systems (POS, ecommerce, CRM, inventory, marketing)
Multiple versions of the truth
No single, connected view of the customer or product

Historically, people have bridged these gaps manually. But AI removes that buffer and exposes every inconsistency in real time.

Data lake ≠ data foundation

A common response to this challenge is to centralise data, typically through a data lake.

But Nimitt was clear that this isn’t enough.

“A data lake helps you store data,” he explained. “It doesn’t make it usable.”

In some cases it can even reinforce the problem, creating a centralised repository of fragmented, passive data.

What retailers actually need is a data foundation.

One that is:

Connected – linking customer, product, inventory and transactions
Governed – with clear ownership, control, and consistency
Real-time – reflecting what’s happening now, not yesterday
Meaningful – enriched with context so AI can interpret and act on it

Because data only becomes valuable when it is understood, not just stored.

From data to intelligence

Strong data foundations are what turn raw information into usable intelligence.

They enable AI to:

Make relevant, accurate decisions
Deliver consistent customer experiences
Operate at speed, without introducing risk

Without that foundation, even the most advanced AI models will struggle to deliver value.

A practical path to AI readiness

One of the most useful takeaways from the session was that building data foundations doesn’t have to mean boiling the ocean.

Instead, Nimitt outlined a pragmatic, phased approach:

Prioritise the problem
Focus on use cases that deliver clear business value
Align your data domains
Fix what matters for those use cases, not everything at once
Establish trust through governance
Ensure data is owned, controlled and reliable
Enable real-time capability
Make data actionable at the moment it’s needed

This shifts data transformation from a long-term project to a scalable capability, one that grows alongside your AI ambitions.

Competitive advantage comes from execution

The retailers pulling ahead aren’t the ones talking most about AI.

They’re the ones executing on data.

With the right foundations in place, organisations can:

Act faster, with confidence
Scale AI beyond pilots
Reduce complexity and cost
Invest with clarity, not assumption

But getting there requires more than internal alignment.

“Transformation starts with the right partner,” Nimitt said, “one that can bring domain expertise, scalable delivery, and contextualised solutions tailored to your existing landscape.”