Mohan Belani is the CEO of e27, a platform dedicated to curating information, connecting stakeholders and fostering the sustainable growth of Southeast Asia’s technology ecosystem.
AI adoption in travel may look impressive on paper, but the results tell a more nuanced story.
Effectiveness: HBX Group found that 65% of travel professionals are already using AI, with 64% saying it improves their day-to-day work. But satisfaction is highest in back-office applications such as fraud detection and expense auditing, while guest-facing chatbots and virtual agents lag, scoring just 3.52 out of five in Business Travel News’ 2026 report. AI is earning trust behind the scenes, but it still has work to do where customers see it.
Depth: Much of today’s AI adoption is a coat of paint applied to existing workflows rather than a fundamental redesign. Amperity found that 80% of travel and airline brands use AI, but only 35% deploy it in the guest experience. Phocuswright found that just 6% of travel businesses are genuinely scaling agentic AI, 22% are beginning to do so, and more than 60% are still experimenting.
Maturity: The trade is splitting into two speeds. OTAs, GDSs and large TMCs are building towards agentic commerce, but only 11% of travel companies can currently complete a real-time booking through an AI agent, according to Skift and Gimmonix. Most SME agents remain in pilot mode.
Challenges: The biggest barriers are less about technology than data and skills. Fragmented customer data limits personalisation, while AI training is often too superficial or left to individuals. In China, 63% of agencies cite data privacy and security as the biggest constraint on wider adoption.
How SME travel agents can rise to the challenge
AI is earning trust behind the scenes, but it still has work to do where customers see it.
Don’t try to build everything at once. Begin with a company-wide exercise to identify the workflows causing the greatest pain points, then choose one or two areas where AI could make a measurable difference. An early win can build confidence and encourage a culture of experimentation and implementation.
Fix the data foundation at the same time. This needs to be done centrally, with data warehousing consolidated to create a single, unified view of the customer.
Training also needs greater investment. It should be treated as dedicated development time, rather than something squeezed into a half-day session or undertaken outside office hours.
To build a moat beyond AI, organisations should lean into what technology cannot do: human judgement in complex situations, higher-value trips and deeply personal experiences.
Finally, get agent-ready by structuring inventory and rates so that AI shopping agents can find and transact with you. Being discoverable in that shift may ultimately matter more than any chatbot on the website.
This article was first published in Travel Weekly Asia’s July-September 2026 issue.Click here to read more from this issue.