As travel adjusts to account for the productivity impact of artificial intelligence (AI), industry teams are figuring out how to keep up.
While Sabre executives discussed the company’s transformation at its Compass ‘26 event, additional speakers attempted to cut through the AI noise and explain “how we can all stay human in a world for machines,” as X. Eyee, CEO of Malo Santo, put it.
Eyee and Isabel Macdonald, senior behavioral scientist for Irrational Labs, outlined how to gain buy-in and how travel companies can manage expectations with evolving technology.
The key takeaway? Humans will shape the industry’s AI future depending on how they present, regulate and interact with the technology.
Gaining AI buy-in
According to Macdonald, teams must actively get on board to implement AI.
The buck doesn’t stop at having a good product and letting people know about it. People need to feel like they are part of the process to value the output.
“It's this fine needle that you need to thread [to make] it really easy for people to feel involved while having the AI really carry the brunt of the effort there.”
Involvement in production also leads to an uptick in trust, but Macdonald cautioned against “pushing people into your AI clothes before they are fully ready.”
“If you are going to default people into an AI experience, you're going to take away whatever the method was that you've had in the past,” Macdonald said. “You want to be sure that you are setting them up for success, because it's going to have this very lasting impact on usage.”
She also emphasized the importance of communicating a “mental model” of AI tools, or how people understand the product. This includes setting specific expectations for what users can do with the product and how they can engage with it.
The temptation might be to communicate to users that your AI product can “do absolutely everything,” but that isn’t likely to help build trust.
Giving users a clear mental model makes them more likely to try specific tasks, while telling them the tool can do anything may lead to more random experimentation.
“You have to think about how your consumers are going to respond and make sure that they actually use their tools,” Macdonald said.
The value of human capital
Beyond managing expectations, travel leaders must also tackle the structural side of AI integration—and that will require a human in the loop.
In the last two decades since Eyee began writing code, the industry has evolved tremendously. It’s moved from old school writing to React applications to cross-platform applications and now to systems that are now capable of building for humans.
“This time last year we were talking about chatbots,” Eyee said. “Now we're talking about replacing entire teams with software.”
But that doesn’t mean that humans will be irrelevant.
“I'm genuinely up here because compliance is really, really hard,” Eyee said. “If your organization does decide to take on building your own agentic infrastructure to do these kinds of transactions from scratch, you're going to have to build in that compliance and governance layer.”
AI should not replace humans but should remove the “machine work” from being human, Eyee continued. It should be used strategically, combining token capital (mathematical representations of words created by AI models) and human capital.
“They're separate things with separate values that are going to be able to align and exist together.”
Humans are also necessary for discernment.
“The humans … shepherd and shape those systems as you go forward, so it's a unique judgment that we bring the relationships, the taste and the accountability that a token calculator can never have,” Eyee said.
Companies are having a difficult time with agentic AI due to the fact that they are using systems in ways they aren’t meant to be used. Maybe the company’s system design isn’t strong, maybe they're pretending hallucinations won’t happen, or perhaps they’re not thinking about cascading failures—the list goes on.
“They're saying we're going to figure that out later,” Eyee said. “They're turning [AI] on in a way that's insecure.”
Governance of AI is critical, and that has to be done by living, breathing humans who need to plan for failures, cost injections and operational issues, according to Eyee.
“We need to allow the technology to fill the role of the token capital while allowing the humans to remain the people who shape the machines that we created instead of looking at it as these machines are capable of doing that,” Eyee said.
*This reporter's attendance at the event was supported by Sabre.