Good data and artificial intelligence go hand-in-hand, says Dan Keto, president and chief technology officer with Easy Metrics, Inc., but companies must make sure that data from multiple sources isn't trapped in functional silos.

In the last 15 years, distribution facilities have experienced a “huge datafication uplift,” Keto says, with more than a dozen distribution data sources to contend with — and, at the same time, less visibility of that data.

Artificial intelligence presents an opportunity to consolidate and make sense of the information. In January of this year, the technology hit an “inflection point,” Keto says, and “exploded” the value of agentic AI. Now, facilities are able to streamline their data-driven processes.

That said, AI is “just a tool,” Keto says. Users need to do more than impose ChatGPT onto their datasets, in an effort to extract and get value out of raw, untransformed data. “That’s not what it was designed for,” he says.

Warehouses seeking to get the most out of AI must begin by building a unified model, one that draws on disparate sources, and aligns the resulting data to the varying needs of each stakeholder — whether they be engineers, finance managers or systems operators. “There’s always going to be compromises between stakeholders as you construct the data model,” Keto says. “Everyone is going to look at the data very differently.”

The primary source of transaction data will be the warehouse management system, but there often will also be a warehouse execution system running the hardware. In the end, all of it comes together to help users calculate and execute on getting the right product to the customer at the right quality and cost, Keto says.