Metcash exploring agentic AI to automate retail ordering

Metcash is exploring the potential for embedding agentic AI into its retail ordering systems as a way to take share from rivals in the wholesale goods distribution market.

The wholesaler, which turns over $19.6 billion per year supplying independent retailers varying in size from recognised brands like IGA right through to small family-run corner stores, has recently switched to conversational search for Sorted, a marketplace to help retailers identify new products to sell.

However, the wholesaler has revealed that its next step with AI could involve giving retail customers access to agentic methods to simplify the process of making routine orders, thereby increasing the potential of securing their orders and loyalty.

Speaking to iTnews at an SAP customer event, Metcash general manager of B2B digital Simon Williams was reluctant to go into details about the wholesaler’s agentic play, citing intellectual property protection considerations.

He did, however, paint a sufficiently detailed picture of how the agentic AI system could work in practice to bring more orders onto Metcash’s books.

“Use an example of one customer that is a small takeaway business. Looking at their data, we can tell that they order on a Monday night at 10.30 pm and they order exactly the same products every single week, so that's a routine for them.

“We know how long they log onto the platform. We know how long it takes for them to order. The thinking's got to go along the lines of … ‘how do we make that job so much easier for them?’

“If we know that the products that they're going to order, how do we bring in an agentic way of working so that that business doesn't have to have to process step every Monday night, because they're not only just using us, because we can see what they're ordering from us and we know what other products they're selling, so they're ordering from elsewhere [as well],” Williams said.

The goal in using agentic AI to make that process simple, Williams confirmed, was to increase the likelihood that the customer will order more from Metcash than its competitors.

Metcash is betting that retailers will be attracted to the quid pro quo of being able to claw back time in their day normally absorbed by dealing with multiple, complex ordering systems.

“They have a myriad of ways that they have to order the products to run their business. It could be some through telephone calls, some through help desk, some online. The way that they work is really complex,” Williams explained.

Williams said that many retail operators had left corporate life to run their own shows but “the efficiency through ordering is not there for them at the moment.”

“How do we drive an agenda of making it easy to do business, making it connected through how they order products, how they potentially sell products, how they work through in terms of their payments that enables them just an easier way of driving business?

"Some of that will be through simple marketplaces where they can procure products, some of it will be through the use of technology like AI that makes sure that we're serving up the relevant products for them to run their businesses,” he said.

Metcash has already made a start, having recently switched to SAP partner Coveo for its AI-driven, semantic search, deploying it for retail customers to use when placing orders from Sorted’s wholesale catalogue.

Williams said that meant Sorted’s search engine is now giving retail customers more relevant results, with products more closely matching their needs higher rankings rather than being pushed to the back.