How Cisco uses AI to find diverse suppliers
The digital communications technology giant employs machine learning vendor Tealbook to gain access to data that fits the types of suppliers it wants.
For years, Cisco Systems' supplier diversity program has focused on hiring firms with a variety of backgrounds.
However, finding suppliers that are owned by women, veterans, and minority groups or operating in poor rural and urban communities, including tribal lands -- what the federal government calls an economic HUBZone -- can be a difficult process if done manually.
To simplify efforts, Cisco turned to Tealbook, a vendor that uses natural language processing to help procurement teams find the right suppliers.
Tealbook uses machine learning methods to scan and ingest public and propriety information from suppliers' websites. The Tealbook system then aggregates and assembles the information for the vendor's customers.
Cisco's use of Tealbook
In 2020, Cisco's supplier diversity team compared Tealbook to other vendors that offer similar services, such as Slate. However, Bryan Wiggins, supplier diversity leader, said Tealbook stood out.
"As a platform, it seemed more proactive than reactive," Wiggins said, adding that other platforms Cisco considered didn't focus enough on diversity as much as Tealbook did.
"As a supplier diversity tool, there was really no comparison in the quality of the diversity information and probably the ease of use," he added.
Wiggins uses the Tealbook platform to determine if a specific supplier is diverse. His team members use the platform's keyword search to find a supplier within a specific parameter that meets the diversity criteria, to doublecheck if a subset of suppliers they have is diverse, and to compare the classifications of diversity that are assigned in Tealbook against Cisco's definition of what is diverse.
"We don't always find a one-on-one match," said Terri Hirahara, business and data analyst at Cisco. This is because Cisco's classification is focused on if a supplier is a minority-owned company or a minority business enterprise. But Tealbook may flag such a company as just a "small, disadvantaged business."
Despite the differences in classification, Tealbook's machine learning technology has proven to be mostly accurate for Wiggins.
"I don't think I've ever found a supplier that Tealbook said was diverse that we later found out was not," he said, acknowledging it's possible that there may have been a supplier that was not listed as diverse by Tealbook but was truly diverse. "There is no perfect golden, magical list."
Bryan WigginsSupplier diversity leader, Cisco
Another challenge with using the platform is sifting through the suppliers that are both diverse and willing to work with Cisco.
"That may actually be an unintended consequence of an AI platform," Wiggins said.
When suppliers register in Cisco's portal, it's clear that they have an interest in working with the communications company. However, since Tealbook's AI technology provides supplier data, Wiggins' team must do some extra work to determine if the supplier wants to work with Cisco.
This means having a conversation with the supplier to determine what their motivation is.
"All the AI in the world is not a substitute for speaking to someone about their about their company and judging someone's motivation to make you happy as a customer," Wiggins said.