Manuel Neto on Why Retail AI Fails Without Purposeful Data

Retail is pouring money into AI while sitting on data it has never truly used. That gap sets up most of the failures to come. Manuel Neto has spent years making organizational data usable inside large, complex businesses, and he treats the problem as cultural before it is technical. He spoke to ClickZ at The Lead Summit 2026 about a single idea. AI only pays off when the data underneath it has a purpose.

Democratized data means delivering it the way each team can act on

Democratization sounds like giving everyone access. Neto frames it differently. “Democratization is when you’re delivering the data in the way they need to be delivered to drive actions,” he said. A creative team and a marketing team need very different things. It is not enough to hand over the numbers. You have to add context, intelligence, and the right format and timing for each group.

Most attempts fail for a simple reason. Companies push the same information across the whole enterprise without matching it to each team. So the effort gets abandoned, it drives no action, and the early momentum drains away. The fix runs through structure. A centralized model handles standardized, enterprise-wide needs. Embedded data scientists then sit inside business units, learn the context, and deliver sharper insight than a central team could alone.

AI built on weak foundations only amplifies what is already broken

Neto sees the ambition gap open at the execution layer. Most companies hold plenty of data, yet they have not activated what they already own. That matters, because AI does not repair a weak base.

“When AI comes in, it’s amplifying broken, it’s amplifying silos, it’s amplifying wrong.”

Instead of helping, it becomes the problem.

By Neto’s estimate, retail AI is roughly 30 percent reality today. “Maybe 60 percent is just a lot of hype,” he said. The capability most large retailers assume they have, and often do not, is data integrity. There is too much data and too little quality assurance. So the order of operations is fixed. Build the structure first, then let AI amplify it.

Purpose comes from tying data to KPIs

Availability is not the real constraint. Most companies already have the numbers. The trouble is that everyone receives the same information regardless of what they do, and purpose gets lost. Purpose means connecting data to the KPIs of the company, the group, and the wider strategy, then building action on top.

A company can distribute numbers across the org, or it can make sure those numbers connect to a KPI and help people act every day. Only the second approach earns the effort. The first looks like progress but then it goes nowhere.

The biggest retail AI value sits in operations

Most AI coverage in retail points at the shopper. It celebrates personalization, recommendations, and search. Neto thinks the larger prize is the systems that feed those customer-facing tools, including inventory control, planning, allocation, and logistics. An AI layer that reviews them end to end and returns the best approach could lift efficiency in ways the storefront cannot. Those systems still need to serve the customer, but they are where the unrealized value hides.

People, not technology, are the hardest part of the change

Ask Neto for the biggest barrier in a legacy retail business, and he does not name a system. He names people, and he means it as a compliment to their importance. “AI requires adoption, it requires change management, it requires literacy in data,” he said. The technology moves exponentially faster than what came before, so teams have to understand the power in their hands. Educate people, and they become advocates.

The human lesson goes deeper for him. He respects where people come from, and he wants room for them to question the data rather than simply receive it. When someone challenges a number, he takes the time to explain it and brings them along. That habit points at his real warning for the industry. “We’re so focused on speed, velocity, and AI enhancements that we’re forgetting the consumers should be at the center of those decisions,” he said. In five years, he expects retail to look back and admit it measured consumer desirability wrong, because it chased velocity and lost sight of the person it was serving.