Hotels and online travel agencies are accelerating AI integration, but trust, data quality, and payment security remain key barriers to fully automated bookings, with industry leaders cautious about the shift from assisted discovery to autonomous transactions.
Hotels and online travel agencies are moving faster to plug their inventory into AI search tools, but the shift from assisted discovery to fully autonomous booking is still limited by trust, data quality and the way accommodation systems are structured. The Travel News report said the industry is working towards so-called agentic commerce, in which AI would handle the journey from search to payment, yet the commercial reality remains more cautious.
That caution is partly technical. According to the report, many hotel property systems, central reservation platforms and customer relationship databases still store information in inconsistent formats, making it difficult for large language models to interpret live room stock, pricing rules and restrictions reliably. Lasse Vinther, managing director of Automation Architects, told Travel News that AI is already useful for research-heavy travel shopping, but said bookings still tend to finish on the supplier’s own site because consumers and businesses want greater certainty before paying.
There is also a structural reason for the delay. Mews said in November 2023 that it had built a direct integration with Expedia Group through its property management system, allowing live updates of rates, restrictions and availability and reducing overbooking risk. The company also said it supports connections to Booking.com, dayuse.com and Google Hotel Search, underlining how the industry is still relying on direct feeds and existing distribution channels rather than fully independent AI checkout flows.
Vinther warned that bias is becoming a second major obstacle. He told Travel News that AI answers can narrow search results to a small set of properties, raising questions over paid placement and preferred partners, particularly as some direct integration deals involve dominant online travel agencies. He added that travel sellers can partly control the output with carefully constructed prompts, but said most providers remain focused on planning and inspiration rather than allowing an AI to complete the transaction itself.
Payments are the hardest step to automate. Travel News said Visa’s Stay Secure study found only 23% of South African consumers would trust an agentic AI to complete a financial transaction, even though more were open to using AI for price comparison. Vinther said Google is furthest ahead in setting standards, while other providers have scaled back ambitions. OpenAI told Skift in March 2026 that travel booking is too complex for AI to manage end to end, and that judgment now appears to define the sector: AI can search, compare and sort options, but the final booking still usually returns to the hotel or OTA.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
