Are AI agents actually removing online travel agencies from bookings?

Rajeev Goswami, CEO of WWStay, argues that AI shopping agents are not removing online travel agencies (OTAs) such as Booking.com and Expedia from hotel bookings; instead, OTAs are becoming the settlement layer those agents depend on. He writes that "AI is removing the storefront, not the infrastructure underneath it." His evidence: independent hotels' RevPAR (revenue per available room) fell 5.4% in 2025 while their OTA dependence climbed to 63.4% of bookings — "worse, not better, in the year AI was supposed to start fixing this." He also points to OpenAI's "quiet retreat from ChatGPT checkout in March" as a tell: even when an agent books through an OTA's API rather than its website, the OTA stays merchant of record and keeps the commission.

Goswami's reasoning centers on where commerce infrastructure sits, not where the user interface sits. He finds that Expedia's B2B segment — which powers other companies' booking tools, including the travel-management-company (TMC) platforms that corporate travel programs route through — grew 24% year-over-year and is now 38% of Expedia's total revenue. Branded hotel chains' OTA channel share, meanwhile, "has barely budged in a decade of 'book direct' campaigns." For Goswami, adding an AI layer on top of a TMC platform changes who makes the booking request, not who collects the commission beneath it; the dashboards travel managers watch "track booking behavior, not who's still collecting the commission underneath it."

This complicates the expectation, voiced elsewhere in the library, that agentic commerce mainly threatens to cut out existing intermediaries: Will brands succeed by building their own AI shopping agents? frames the risk as one of being bypassed by an agent layer, but Goswami's travel-industry case shows a different outcome for a different class of intermediary — the OTA absorbs the agent layer rather than being displaced by it. It also parallels the mechanism in Why do consumers delay adopting agentic search?: both describe incumbent platforms retaining structural advantage under agentic commerce even when the surface experience seems to favor the consumer or a new entrant. It echoes Is non-human traffic really half of all internet use? in showing an existing commercial layer adapting to, rather than being erased by, agent-mediated access.

Goswami's piece is a subscriber brief built from his own review of 2026 industry figures and a single discovery-call anecdote, not an independent audit of OTA contracts or TMC platform agreements; he states plainly that "nobody's actually measuring whether an AI layer on a TMC platform changes the underlying OTA-supply relationship," leaving his claim about embedded dependency as a question he poses rather than a figure he can cite. The data he does cite — RevPAR, OTA booking share, Expedia's segment revenue — describes the hotel and TMC channel specifically, so it should not be read as a general claim about agentic-commerce disintermediation outside travel.