Businesses all over the world, both big and small, are grappling with the task of integrating rapidly advancing AI technologies into their operations. Brex, a corporate credit card company, has also found itself wrestling with this issue. By switching gears and completely rethinking their procurement approach, Brex has ensured that it remains in step with the frenetic pace of technology.

The challenge with AI adoption was discussed in depth by Brex's CTO, James Reggio, during the HumanX AI conference this past March. Initially, Brex tried to pin down suitable AI tools using its standard procurement strategy, but quickly realized that traditional months-long trial periods weren't going to cut it.

Reggio explained that due to the lengthy nature of the procurement process, teams would lose interest in the requested tool by the time it completed all necessary internal checks. This realization led Brex to overhaul its procurement process.

The remodel started with creating a new framework for data processing agreements and legal validations. The new structure sped up the evaluation of potential AI tools and their subsequent distribution amongst the testers.

To identify worthwhile investments, Brex developed a "superhuman product-market-fit test." This method assigns a pivotal role to employees in selecting tools depending on where they find the most value.

The strategy involves identifying the employees getting the maximum benefit from a tool and evaluating if it has sufficient uniqueness for retention. The approach has resulted in the deployment of around 1,000 AI tools at Brex over the past two years, with some of them not being renewed later.

As a part of the new procurement arrangement, every engineer at Brex receives a monthly budget of $50 for licensing software tools from an approved list. This delegation empowers individuals to make decisions that optimize their workflows, which has underscored the decision to facilitate experimentation with diverse tools.

This revised strategy has also enabled Brex to determine the necessity of broader licensing agreements for certain software by accurately demarcating the usage patterns among engineers.

To sum up, Reggio advises businesses to embrace the anarchy that comes with the adoption of AI tools. There will be bumps along the way, and that's perfectly fine. The risk of being left behind is greatly amplified if one fixates on making the right decision from the get-go. It's essential not to stall for months evaluating everything minutely before deployment as future unpredictability constitutes the very nature of the tech world.