How Brex Adapts to AI by Embracing Flexibility
Many companies struggle to adopt AI tools as the technology evolves faster than traditional procurement cycles. Corporate credit card startup Brex faced similar challenges but found a solution by overhauling its approach to software evaluation.
At the HumanX AI conference, Brex CTO James Reggio shared how the company initially used its standard procurement process for AI tools, only to realize it was too slow. “The procurement process itself would run so long that teams lost interest in the tool by the time we got through internal controls,” Reggio explained.
To address this, Brex developed a new framework for data processing agreements and legal validations, speeding up the vetting process. The company also introduced a “superhuman product-market-fit test” to determine which tools delivered real value beyond pilot phases.
Employees now play a key role in tool selection, with engineers receiving a $50 monthly budget to license approved software. “By delegating spending authority to individuals, they make optimal decisions for their workflows,” Reggio noted. This decentralized approach has helped Brex identify which tools merit broader licensing.
Reggio advises enterprises to “embrace the messiness” of AI adoption, accepting that not every decision will be perfect. “Overthinking evaluations for months risks falling behind in a fast-changing landscape,” he warned.
