Brex's AI Strategy: Embrace the Messiness to Stay Ahead
It's no secret that keeping up with the rapid evolution of AI is a challenge for any company. Even Brex, a corporate credit card company, had to rethink its strategy to avoid falling behind. I mean, who hasn't felt overwhelmed by the sheer number of new AI tools popping up every day?
Initially, Brex tried to use its standard procurement process to evaluate AI software. But according to Brex CTO James Reggio, that months-long piloting process simply wasn't fast enough. By the time they'd finished assessing a tool, the teams that had requested it had already lost interest. I can totally relate – it's like trying to assemble furniture with instructions written in hieroglyphics!
So, Brex decided to shake things up. They created a new framework for data processing agreements and legal validations, which sped up the vetting process. This allowed them to get potential AI tools into the hands of testers much faster. They also implemented what they call a "superhuman product-market-fit test". It sounds fancy, but it basically means giving employees a bigger say in which tools the company adopts. Instead of top-down choices, they empower people to use what adds value to their workflows.
To add to that, engineers at Brex get a monthly budget to spend on software tools from a pre-approved list. This way, the people actually using the tools get to make the call on which ones are best for them. It's like giving them the keys to the candy store, but with a responsible budget, of course.
What I found particularly interesting is that Brex isn't afraid to cut its losses. Reggio mentioned that they've already canceled or not renewed several larger deployments of AI tools. It just goes to show that not every tool is a winner, and it's important to be willing to move on. Because, let's be honest, in the world of AI, things change in the blink of an eye. As Reggio said, the key is to "embrace the messiness" and accept that you won't always make the right decision at first.
Taking months to carefully evaluate every single option might seem like the responsible thing to do, but you might just miss the boat entirely. In short, it's better to dive in, experiment, and learn as you go, than to get paralyzed by analysis.
Source: TechCrunch