AI in Retail - Beyond the Pilots

Artificial Intelligence is everywhere in retail headlines. Analysts estimate AI could create $240–$390 billion in value for retailers globally — equal to a 1.2–1.9 percentage point margin uplift across the industry.

But here’s the catch: a recent study found that 95% of generative AI pilots fail to deliver measurable business impact.

So the question isn’t whether AI has potential. It’s whether retailers can turn that potential into results.

‍

Why the Hard Part Isn’t the AI

Retailers already generate vast data streams:

• POS transactions
• Ecommerce behaviors
• Loyalty programs
• Inventory and supply chain movements

Yet these systems often sit in silos. In fact, only 14% of retailers report having real-time, unified customer data across channels.

That’s why so many AI pilots stall. The models themselves are powerful, but the inputs are fragmented, inconsistent, or outdated. AI is only as good as the fuel it runs on.

‍

Managing Change is Just as Critical

Even with good data, adoption is often the stumbling block.

• In a 2024 survey, 3 in 5 executives said quality and reliability of AI output is their top concern.
• 56% of retailers have increased AI budgets, but most admit scaling beyond pilots is their biggest challenge.

The lesson: AI success isn’t about one great model. It’s about disciplined program management — setting clear checkpoints, building trust with associates, and measuring ROI incrementally.

‍

Where AI is Already Delivering

Despite the failures, there are real wins:

• Walmart: rolled out AI task management and translation tools to 1.5 million associates, boosting efficiency and inclusivity.
• Tesco: testing “VAR-style” AI video review at self-checkout to reduce shrink — tackling a $100B+ annual loss problem in U.S. retail alone.
• Instacart: “Ask Instacart” uses AI to help customers find items and plan meals.
• Walmart & Sam’s Club: using AI to power retail media networks, now one of the fastest-growing profit pools in retail.

Each example shows the same truth: success comes not from the algorithm alone, but from how AI is embedded in everyday workflows.

‍

The Road Ahead

The retailers who get this right will:

• Connect their data across legacy and cloud systems.
• Apply project discipline, moving from proof-of-concept to production.
• Establish governance, ensuring accuracy, transparency, and trust.
• Focus on adoption, empowering associates and customers alike.

That’s how innovation becomes impact.

‍

A Perspective from Makira

At Makira, our strength lies in helping retailers turn complex technology into measurable outcomes.

From integrating systems and building reliable data pipelines, to applying disciplined project governance and ensuring adoption at scale, we specialize in making new technology — including AI — work in the real world. Because whether it’s POS, ecommerce, or now AI, the hard part is the same: connecting the dots, managing the change, and keeping projects on track.

AI may be the newest tool in retail’s kit, but the foundations that make it succeed are the same foundations we’ve been helping retailers build for years.

Closing thought: In retail, the true value of AI won’t be in pilots or proofs of concept. It will be in the projects that make it to scale — the ones that connect data, empower people, and deliver measurable outcomes.

And that’s where the opportunity lies.

‍