Expedia cuts 500 jobs to fund AI hiring as it opens sales to AI agents

Expedia Is Selling to AI Agents. Its Own Annual Report Asks Whether That Was the Right Call.

Expedia Group has started selling its travel inventory to a new kind of customer: AI agents. The company quietly launched a business-to-agent, or B2A, function this month, restructuring how hotels, flights, and packages are presented so that machine reasoning systems — not human travelers — can read, evaluate, and act on them, according to Skift. The announcement landed as a product story. Buried in the same news cycle was the reason the build makes economic sense.

Three months earlier, on a public earnings call, CFO Geoff Schenkel told investors that Expedia had cut 3 percent of its workforce — roughly 500 people — and that the savings would fund hiring in AI and machine learning, GeekWire reported. The framing was operational. The implication was not: Expedia is simultaneously reducing its human operational footprint and investing in the infrastructure that makes its human workforce less necessary. The B2A announcement is the product of that investment.

The timing is not accidental. AI agents do not respond to brand the way human travelers do. "Brand recognition, a cognitive shortcut for human buying decisions, means nothing to AI agents that can reason through every option every time," Skift noted. Structuring inventory for agentic discovery requires different data practices than structuring it for human search — more granular, more machine-readable, more continuously updated. Expedia's B2A function is doing that work.

Expedia's annual report, filed with the SEC in January, provides the company's most explicit acknowledgment that the transition carries risk. For the first time, the 10-K names generative and agentic AI as a competitive threat: the "rapid emergence and adoption of generative and agentic AI is likely to further intensify competition." The phrasing is careful — no confession of obsolescence — but the admission that the technology Expedia is investing in may intensify competition rather than insulate it is notable in a document that typically frames competitive risk in terms of other OTAs and meta-search aggregators.

The contradiction — investing in the infrastructure that agents read, while warning that agents may make Expedia less competitive — is visible in the same SEC filing. The 10-K also flagged that "agentic booking capabilities that may lack strong consent controls could significantly increase fraud risks," GeekWire noted. If an agent can book travel on behalf of a user without clear authorization protocols, the financial and operational exposure grows. This is a cost of the transition, separate from the competitive threat.

Morgan Stanley's read of the same moment is more benign for online travel agencies. "Early agentic AI tools appear more likely to drive users toward OTA platforms for booking rather than replacing them with direct in-AI transactions," the bank argued recently. The reasoning: agents need comprehensive, structured data to work with, and OTAs hold more inventory in more consistent formats than any single airline or hotel. In this framing, Expedia's B2A build is a defensive position, not a risky bet.

The structural counterargument is harder to dismiss. Every booking routed through an agent is a transaction where Expedia's brand investment did not do the persuasive work. The marketing spend that has historically sustained Expedia's competitive position is built for human decision-making. The B2A infrastructure is built for something else. The 10-K language suggests Expedia's own legal and finance teams see the distinction.

The company's recent results show why the transition is happening now rather than later. Q4 gross bookings and revenue grew 11 percent year over year, per GeekWire. B2B revenue — the API and affiliate channels through which agents increasingly interface with travel inventory — grew 25 percent in the first quarter, versus 8 percent for the consumer business, Skift separately reported. The base sizes differ. The growth differential does not prove an agentic transition is underway. It is consistent with one beginning.

What is unusual is that Expedia has described the transition in public, on a earnings call, in terms that made the labor replacement explicit. Most companies running this transition are not simultaneously telling investors they are cutting headcount to fund the automation that makes those headcount reductions possible. The B2A announcement is the product. The February call is the subtext. Together they describe a company that sees the agentic transition as both inevitable and ambivalent about its own role in it.

The question the 10-K raises — and does not answer — is whether Expedia's bet on legibility pays off or accelerates its own displacement. Being readable to agents is an advantage as long as those agents are working for Expedia. The moment they work against it, the infrastructure is in place for the competition to use.