Cutting Through the Noise: Inside the Retail AI Council's Latest Council Collab

Count the number of "AI-powered" pitches in your inbox right now. Got a number? Good. Now ask yourself how many of those products are doing something that wasn't already possible three years ago with a solid data team and some well-structured queries. 

That's not a knock on AI. It's a sorting problem. And sorting real intelligence from good marketing was the driving question behind the Retail AI Council's latest Council Collab, a roundtable featuring Andy Laudato (COO, The Vitamin Shoppe), Steve Williams (CTO, Buff City Soap), Ron Cason (Ashley Furniture), and Jeff Goethals (CIO, Centric Brands), moderated by Joe Dela Cruz. The conversation was unscripted and rooted in the kind of operational experience you can't fake. Here's what stood out. 

The "AI-Washed" Filter 

The panel gave a name to something a lot of retail leaders have been feeling: "AI-washed." It's when the intelligence layer on a product is more cosmetic than functional. The branding says AI, but the workflow says otherwise. 

Their reframe is worth borrowing: stop asking "Does this product use AI?" and start asking "Which decisions does it make better, and with what data?" Can it reason through a pricing exception? Flag a promotion conflict before it hits the floor? Route inventory in a way that takes real work off someone's plate? 

It's a higher bar. But it's also a fairer one. Because when you evaluate that way, the genuinely strong solutions don't have to compete with noise. They just have to show their work. 

"D-Platforming" Over Re-Platforming 

Not the social media kind. 

The traditional re-platforming story goes like this: rip out a monolithic system, spend two years and a serious budget replacing it with another monolithic system, cross your fingers. The panel sees that model fading fast. What's taking its place is more modular - price optimization as its own service, inventory intelligence as a separate layer, promotion management decoupled from the rest of the stack. 

The advantage is both technical and strategic. When your pricing engine isn't hardwired to your order management system, you can actually test things. Try a new AI capability in a contained way. See what it does before you commit. The brands pulling this off, the panel noted, tend to have disciplined architecture rather than the biggest checkbooks. 

For solution providers, this should be welcome news. Modular buyers have a lower barrier to trial. They're more willing to try new tools. The flip side? Your product needs to stand on its own, not just as part of a bundled suite. 

Data First. Then Intelligence. 

Nobody loves talking about data hygiene. It's the flossing of enterprise technology. But the panel made a case for it that was hard to argue with. 

AI tools are only as useful as the data feeding them. Product taxonomies with inconsistencies, customer records with duplicates, inventory counts that don't quite line up - none of that is unusual, but it all affects what AI can realistically deliver. Layer intelligence on top of messy data and you don't get smarter operations. You get confident-sounding wrong answers, faster. 

The panel framed this as a sequencing issue rather than a readiness issue. It's not that these organizations can't adopt AI. It's that the ones who clean up the data layer first get dramatically better results when they do. 

For IT leaders, that sometimes means being the person who pumps the brakes. Not to kill momentum, but to protect it. 

Where "Close Enough" Doesn't Cut It 

Not all use cases carry the same stakes. The panel spent time on this, and the distinction matters. 

Product recommendations, demand forecasting, ad targeting - these are areas where AI can operate with some margin for error and still deliver strong results. But in regulated spaces like health and wellness, that margin disappears. Dosage information, ingredient claims, compliance data - those need to be right, full stop. A hallucination in a product recommendation might mean a weird suggestion. A hallucination on a supplement label is a different category of problem. 

The point wasn't to wall off entire industries from AI. It was to be specific about matching the right type of AI to the right use case. That precision is good for retailers. And it's good for solution partners too, because it means the conversation starts with fit rather than hype. 

PoCs as a Trust-Building Tool 

This might be the most immediately useful idea from the session. 

There's a familiar pattern in enterprise IT that the panel wants to break. A vendor delivers a strong out-of-the-box solution. Then the buying organization spends months customizing it until the original product is barely recognizable. Everyone's frustrated. The budget's gone. The thing still doesn't feel right. 

The alternative: let teams use the tool as-is before customization even enters the conversation. Run a real proof of concept. See where the product genuinely falls short versus where the discomfort is just "this doesn't look like what we had before." Those are two very different problems, and you can't tell them apart without living with the tool first. 

Paired with a modular stack, this approach becomes low-risk and high-signal. Teams build real buy-in because they've experienced the product, not just seen a demo. And vendors benefit too - you end up with a customer who understands what they bought, not one who's already planning to rebuild it on day one. 

What It Adds Up To 

The leaders on this panel aren't sitting on the sidelines. They're actively building AI into their operations. But they're doing it with the kind of rigor that comes from having been through a few hype cycles before - clean data first, modular architecture, clear evaluation criteria, and the discipline to test before they scale. 

For the retail AI ecosystem, this is a good thing. Sharper buyers don't slow down adoption. They speed it up. Because when a solution earns their confidence, it doesn't just get a pilot. It gets a rollout. 

The Retail AI Council's Council Collab series brings together retail technology leaders for candid, practitioner-led conversations about AI in retail. Learn more at retailaicouncil.com.