Can you really trust AI with procurement tasks?

Once viewed as a dusty back-office function, procurement is now regarded as a strategic priority with significant impacts on both top and bottom lines. As a complex, data-intensive process that typically involves large numbers of suppliers, products and services, it's also uniquely suited to the analytical and communication capabilities of generative AI, as well as the workflow automation and decision-making of agentic AI.

Many of the tasks handled by procurement workers, from choosing the right suppliers to communicating specifications and tracking price lists as well as delivery dates, can now be largely automated with AI.

To that end, Globality, based in Palo Alto, Calif., spent more than a decade training the AI data models in its cloud-based procurement platform. An AI agent named Glo ranges across the platform, applying what Globality calls "deep procurement expertise and broad contextual understanding" to respond to requests, coordinate supplier communications, make recommendations and execute transactions.

In this episode of Enterprise Apps Unpacked, Globality CTO Keith McFarlane explains how the software works, where it fits into the procurement process and ERP ecosystem, and how Glo works harmoniously alongside humans in the loop.

Learning to trust AI

Keith McFarlane

Given the huge operational and financial stakes of procurement, this level of automation will only work if people trust the AI not to make mistakes, according to McFarlane.

"If you get the team to trust the outputs of the AI -- if the agentic system is producing the results that it should -- then you're going to see higher utilization," he said. "At first, it takes some handholding, in many cases. You have to get folks as comfortable as possible with using the tool as it's meant to be used while letting go of certain elements of the tasks that they would normally do day-to-day."

Other topics discussed in the podcast include the following:

• Whether Globality's AI automation threatens procurement jobs.
• How much training Glo needs from users and what it can learn on its own.
• Where potential cost savings come from.
• How procurement jobs might evolve in the near future if agentic AI continues to improve.

David Essex is an industry editor who covers enterprise applications, emerging technology and market trends, and creates in-depth content for several TechTarget websites.