Companies have found it challenging to adopt suitable AI tools because the technology is advancing much faster than their typically lengthy sales cycles.
Brex, the corporate credit card startup, faced the same challenge as larger enterprises. To avoid falling behind, it overhauled its software procurement strategy entirely.
Brex Learns Its Traditional Procurement Process Can’t Keep Up with AI’s Rapid Pace
At the HumanX AI conference in March, Brex CTO James Reggio told TechCrunch that the company initially attempted to evaluate AI tools using its standard procurement process. However, it soon realized that its months-long pilot approach was unworkable.
“In the year after ChatGPT launched, so many new tools were emerging that our procurement process ended up taking so long, the teams requesting the tools often lost interest before we even cleared all the internal approvals,” Reggio explained.
That’s when Brex recognized the need to overhaul its entire procurement approach.
Streamlining AI Adoption
According to Reggio, the company began by creating a new framework for data processing agreements and legal reviews specifically for AI tools. This streamlined process enabled Brex to evaluate tools more efficiently and put them in the hands of testers much faster.
He added that Brex now uses a “superhuman product-market-fit test” to determine which tools are worth investing in after the pilot phase. This method gives employees a greater voice in selecting tools, based on where they’re actually seeing value.
“We closely engage with the people getting the most out of each tool to assess whether it’s truly valuable enough to keep,” Reggio explained. “We’re now about two years into this new phase, with around 1,000 AI tools used across the company. We’ve definitely ended or chosen not to renew at least five to ten major deployments.”
Brex also gives its engineers a $50 monthly budget to license any software they choose from a pre-approved list.
Empowering Engineers to Choose Their Own Tools Drives Workflow Optimization and Diversity
“By giving spending power to the individuals actually using the tools, they can make the best choices for improving their workflows,” Reggio said. “What’s interesting is that we haven’t seen everyone gravitate toward the same tools. That reinforces our belief in making it easy to test a wide variety—people aren’t just defaulting to one popular option like Cursor.”
This strategy has also helped Brex identify when broader software licensing deals are needed, thanks to a clearer picture of how many engineers are actually using each tool.
Overall, Reggio believes the best way for enterprises to navigate the rapid pace of AI innovation is to “embrace the messiness” and accept that the process of choosing the right tools will be imperfect—and that’s completely fine.
“The key is recognizing that you won’t always get it right on the first try,” Reggio said. “What matters most is avoiding paralysis. The biggest mistake would be overanalyzing and spending six to nine months evaluating every option before making a move—because by then, the landscape could be completely different.”
Read the original article on: TechCrunch
Read more: X Limited Access To Reuters Accounts Within India