In today’s fast-moving AI landscape, waiting around simply isn’t an option. Brex realised that clinging to slow, traditional procurement methods meant missing out on the latest tech trends. If you’ve ever struggled with long approval processes, you’ll appreciate how this new strategy cuts out the delays and empowers teams to act quickly.
At this year’s HumanX AI conference, CTO James Reggio explained how the company’s old process couldn’t keep up with the surge of new tools following ChatGPT’s rise. By the time a tool navigated through all the internal controls, its initial excitement had faded. Recognising this, Brex overhauled its approach—creating a streamlined framework for data processing agreements and legal validations that lets them assess potential AI tools with much greater speed.
One standout element is what Reggio calls a “superhuman product-market-fit test.” What this means is that those on the front lines—the employees who actually use these tools—play a key role in deciding which software is unique and genuinely useful. This hands-on method has already led to testing roughly 1,000 AI tools, with a few larger-scale deployments being paused when they didn’t quite hit the mark.
To keep innovation at the heart of the process, Brex has granted each engineer a monthly $50 budget to license software from a pre-approved list. By putting spending power in the hands of the people who know what they need, the company ensures that decisions are both practical and aligned with daily workflows. When the data shows that more extensive licensing might be beneficial, they know exactly when to pivot.
This flexible, no-nonsense approach isn’t about striving for perfection from the outset—it’s about making timely decisions in a rapidly changing world. Instead of spending months deliberating every move, Brex opts to act quickly, trusting that recalibrations can be made as the landscape evolves. After all, waiting too long might mean missing the opportunity to lead in innovation.
