Successful AI in retail only works when operations change with it

For many retailers, artificial intelligence has reached an inflection point, not because the technology has suddenly matured, but because expectations have. After years of pilots, dashboards, and experimentation, supply chain leaders are increasingly asking a harder question: What business outcome does this actually change?

According to Andrea Morgan‑Vandome, SVP and chief innovation officer at Blue Yonder, that shift in mindset is the single biggest determinant of whether AI delivers value or simply produces more noise.

“One of the things we have found that is resonating really well is that … retailers have to change their operational procedures,” Morgan-Vandome told Supply Chain Management Review. “Otherwise you’re using AI to get more bad data.”

AI doesn’t fix broken processes

That distinction shows up clearly in areas like returns, where AI is often positioned as an optimization tool without addressing the underlying workflow.

Retailers, Morgan-Vandome explained, are increasingly rethinking how returns are handled, not just how quickly they’re processed. The goal is to move returned goods back into sellable inventory faster, reduce transportation dwell time, and capture more granular data about why products are coming back in the first place.

“As you do those returns and now that you know why it is being returned that data can be used to improve processes,” she said.

In other words, AI only becomes valuable when it is paired with operational change. Knowing why an item is returned is only useful if that insight flows back into planning, merchandising, fulfillment, or transportation decisions, she said.

From insights to execution

Morgan-Vandome emphasized that AI’s real value emerges when it connects planning and execution, rather than living in an analytical silo.

“AI is also an enabler of helping people bring together end-to-end processes. And you can drive big value out of it,” she said.

One example involved a grocery retailer trying to reduce waste while ensuring products were still on shelves when customers needed them. Solving that problem required more than better forecasts, it demanded tighter coordination between demand sensing, replenishment, and execution.

The same principle applies inside the warehouse. Layout decisions, traditionally revisited only during periodic redesigns, can now be continuously evaluated.

“In a warehouse, the issue is how to lay out the warehouse,” Morgan-Vandome said. “Previously that was done at specific time periods. With an AI agent, you can start to look at how to lay out the warehouse in the most efficient way and the most efficient way may be store-ready pallets.”

The key, she noted, is combining machine learning with agents that can execute on insights that are useful for people, not simply surface recommendations that never make it into operations.

Anchoring AI to outcomes

One of the challenges Morgan-Vandomen noted is “AI-for-AI’s-sake” initiatives.

“We actually anchor on how will we change the outcomes of the business,” she said. “Because if what we are doing doesn’t actually change the outcome, why would you do it?”

That philosophy extends to how Blue Yonder works with customers. While some companies still arrive with a predefined “AI budget,” Morgan-Vandome said most conversations now start with a business problem, often one that turns out to be broader than initially described.

“Normally, they are coming to us with a problem,” she said. “And sometimes when we dig into it, the business problem they present is just part of a bigger problem.”

A retailer struggling with out-of-stocks, for example, may believe it has an inventory issue. But the root cause could be ordering logic, promotion timing, supplier constraints, or even marketing misalignment. The challenge, she said, is phasing the solution correctly, deciding what to fix first and determining how quickly value can be delivered.

Speed matters more than perfection

That focus on outcomes has also reshaped how retailers think about transformation timelines.

“There are those big, long transformations that take a lot of time,” Morgan-Vandome said. “But people are also looking at how can you drive outcomes quickly.”

Rather than waiting for perfect data or fully modernized systems, many retailers are using AI to surface where data quality actually matters most. Proof-of-concept projects, she said, often reveal which inputs are truly constraining performance and which aren’t.

In one recent three-month test, a retailer lifted sales and reduced out-of-stocks using the same amount of inventory, simply by changing how decisions were made.

“That’s where AI helps prioritize what to fix,” she said.

Embedded AI, not AI features

A critical distinction in Morgan-Vandome’s view is that the most effective AI is often invisible.

“Our solutions are built from the bottom up with AI,” she said. “A lot of the time AI is part of the solution, but it’s embedded so it’s not an ‘AI solution.’”

That approach contrasts with tools that pull data out of operational systems, analyze it elsewhere, and then push recommendations back. Instead, AI becomes part of the workflow itself, shaping decisions as they happen.

“There is a lot in the market where people take data out, put it aside, and let the AI make decisions,” she said. “Where we make it part of the solution is how we apply it to change operational procedures and give it to planning.”

From innovation to adoption

Morgan-Vandome’s role spans both innovation and supply chain advisory, reflecting a broader shift inside Blue Yonder toward adoption, not just invention.

“It’s really a switch around adoption,” she said.

That includes organizational changes on Blue Yonder’s side, with teams staying engaged through implementation to ensure customers understand not just how to use the technology, but why it exists in the first place.

“Everything we are doing,” she said, “we are doing to solve a problem rather than just sitting there with an agent to ask questions.”

Ultimately, AI does not replace strategy, process design, or operational discipline, it amplifies them. When paired with clear outcomes and real process change, it can drive meaningful results. Without that foundation, it simply accelerates existing inefficiencies.

SC
MR