With less than 60 days until the GenAI-themed Innovation Day, Birnbaum and the entire technology organization wanted to empower teams to be able to test out different GenAI use cases independently and securely. United had partnered with Slalom to build a machine learning operations (MLOps) platform that streamlines the creation, deployment, and management of machine learning models, so United’s leadership knew Slalom had the knowledge and expertise to help extend that platform to accommodate GenAI.

We got to building and integrating a GenAI “testbed” that securely leverages United’s data infrastructure. Built on AWS, the testbed uses Amazon Bedrock to give teams access to a variety of proprietary and open-source GenAI models including Amazon Titan, Anthropic’s Claude, and Meta’s Llama. Many participating teams used the testbed to determine which AI model best fit their use case in time to demonstrate that use case at the Innovation Day. In total, executives across the enterprise, including the CEO of United, evaluated more than 30 use cases as part of the Innovation Day. 

Grant Milstead, a VP of Digital Technology at United who helped lead the preparations for the GenAI Innovation Day, emphasized how important it was—and is—for United’s teams to be able to access the testbed on their own. “At one point we had, like, 30 teams doing different ideas,” said Milstead. “With that many hands in the cookie jar, it couldn’t just be one person at Slalom giving access out and showing how this worked. We had to be able to get a lot of developers in there and it had to be something that was kind of enterprise-grade versus just a science project.”

The GenAI Innovation Day was a huge success, and the expanded MLOps platform will support United teams’ innovation and experimentation well into the future—and far beyond 30 use cases. Already, two use cases showcased at the Innovation Day have been released as solutions with Slalom’s help. 

United refers to the first solution as an enhanced version of its “Every Flight Has a Story” (EFHaS) messaging program. Even before the arrival of GenAI, a team of “Storytellers” at United have informed travelers about unexpected changes to flight statuses, like delays, through detailed but easy-to-understand custom messages. With the new GenAI capabilities of EFHaS, Storytellers will be able to customize 50% of all flight updates United sends, up from 15%. Storytellers still carefully review each AI-generated message but can ultimately customize and share a much greater volume of messages than they could have prepared manually. The result is more travelers receiving top-notch, transparent updates as soon as their flight statuses change. 

The second solution is a customer care copilot that leverages GenAI to help agents provide faster and more empathetic responses to the thousands of email inquiries they receive from customers every day. 

In travel, plans change. United’s initial GenAI use cases give employees the grace to be ready for this—and to be ready to help customers navigate it. We think our work with United proves that in the right hands, artificial intelligence can enhance emotional intelligence. It can help people do what only humans can: care. 

”We have the greatest employees in the industry, in my opinion,” said Stacey Wronkowski, a VP of Digital Technology at United whose team leads its GenAI Center of Excellence (CoE), a dedicated team assembled to guide and scale United’s GenAI vision. “We are transparent people. We are caring and empathetic people. What it comes down to is, we don't have the ability to handcraft a personal message to every customer, given our size. That connection with customers and messaging to them that is personal and relevant would be hard to scale without a capability like this.”

Given how fast GenAI is evolving, it’s almost impossible to imagine how United might be using the technology a year or even a quarter from now—almost. With a robust platform at its fingertips, a talented CoE in the pilot’s seat, and our continued partnership, United’s GenAI future is clearer than ever. But rather than speculate here about what use cases could be just around the corner, we’ll spend our last words on something a little more human by thanking United’s teams for the opportunity to join the first “flights” of their GenAI journey. 

On behalf of Slalom and the entire crew, thank you for choosing to fly with us, and welcome to the age of AI.