SHOW NEWS

*Insights from the AI in Action panel and BTN's global AI research, presented at Business Travel Show Asia Pacific 2026.*

Artificial intelligence dominated the agenda at Business Travel Show Asia Pacific 2026. Across two standout sessions — the "AI in Action" panel featuring buyers, TMCs and technology providers, and BTN Group's exclusive preview of its global AI research — a consistent picture emerged of where AI is genuinely delivering value in managed travel today, and where the hype outpaces reality. Here are the five lessons every travel leader should take away.

1. AI adoption is widespread — but still shallow

BTN Group's survey of 293 travel buyers, previewed by Vice President of Content Elizabeth West, found that around three-quarters of organisations are using AI in their travel programmes in some form. But depth is another story: just 8% have embedded AI across multiple travel workflows, and only 1% say AI plays a genuinely strategic role in their programme.

Asia Pacific buyers are slightly faster out of the gate — only around 15% of APAC-remit buyers report using no AI at all, versus 25% globally — but maturity levels are almost identical worldwide. Momentum, however, is building: 18% of buyers expect AI use to increase dramatically in the next 12 months, with a further 69% expecting a gradual ramp-up. As West cautioned, drawing on three years of Wharton School research on enterprise AI adoption, progress can feel gradual right up until the fundamentals are in place — and then it accelerates fast. Complacency is the real risk.

2. The highest-ROI AI applications are the ones travellers never see

Here's the counterintuitive finding from BTN's research: high usage does not equal high value. Traveller-facing chatbots are the most widely deployed AI application, used by 52% of surveyed buyers — yet they earned one of the lowest effectiveness ratings, at 3.52 out of 5. AI-assisted reporting, used by 44%, fared similarly.

Compare that with expense auditing and fraud detection: adopted by fewer than 30% of organisations, but rated 4.07 out of 5 for effectiveness — the highest score in the survey. RFP support, sourcing, forecasting and supplier performance analysis showed the same pattern: low adoption, high impact.

The panel's real-world numbers back this up. Marten Jagers, Vice President for Asia Pacific and Japan at Emburse, told the audience that clients applying AI at the front of the travel and expense workflow — capturing and validating receipts, invoices and compliance checks before an expense report is even submitted — are seeing up to 70% less time spent on the T&E process and compliance improvements of upwards of 90%.

The common thread, as West noted, is that the most effective applications work on data you already have rather than trying to change traveller behaviour. If you want a quick, defensible win, start there.

3. Governance and data ownership are the new battleground

If GDPR and cookie compliance felt painful, "you need to be ready for AI compliance," warned Apurva Goswami, CTO for Asia Pacific at Corporate Travel Management (CTM) — pointing to the EU's AI regulation, due to be fully formalised by August 2026, and China as the most stringent regimes.

Yet BTN's research found only 29% of organisations have a formal AI policy covering travel, and nearly half have none at all. More concerning: more than half of buyers said they were comfortable simply accepting their suppliers' AI deployments as-is, only 16% expressed any concern about how suppliers apply AI to their data — and just 2% treat a supplier's AI roadmap as a major differentiator in RFPs.

Accenture offers a model for doing this properly. Thomas Lempart, who manages buying channels within one of the world's largest travel programmes — roughly 350,000 active travellers and close to $2 billion in annual spend — explained that Accenture contracts every supplier directly, and any tool with "even a shade of AI" triggers a full legal and IT review. Deployment stops until suppliers complete testing and sign a dedicated AI schedule of more than 15 pages covering data storage, privacy, model training, whether client data trains other models, and auditability of AI decisions.

The essential questions every buyer should ask: Who owns the data? (The client, both Emburse and CTM confirmed — the supplier owns only the technology IP, and under GDPR terminology the client is the controller, the TMC the processor.) How was the model trained? Is my data training other models? Can I audit why the AI made a specific decision?

4. AI works best when it doesn't force travellers to change their behaviour

Lempart offered a simple test for separating useful AI from hype: does it require travellers to drastically change how they behave? Accenture's AI investments focus on cost, leakage and "nudging" — steering travellers towards compliant, cost-effective choices inside the tools they already use, whether that's the online booking tool or mobile notifications. "You have to remove the friction," he said. "Everything has to happen in one place." Success is measured through freed-up time, hotel attachment rates and cost, not novelty.

Disruption management is where this principle shines. Goswami shared a striking case study: when Cyclone Alfred hit Queensland in March 2025, CTM's AI-powered assistant Scout absorbed a 343% surge in cancellation requests and a 150% increase in booking changes without errors — and without CTM needing to ramp up thousands of consultants. Travellers made changes in under two minutes instead of waiting on the phone, and CTM reports a 20–30% uplift in customer service productivity overall. Notably, CTM frames the benefit as scaling service quality, not cutting headcount.

5. AI won't replace travel managers — it will make the best ones more strategic

Only 4% of buyers in BTN's survey said AI has created serious concern about their long-term relevance. Far more telling: 16% said AI has already made their role more strategic, and 38% report efficiency and effectiveness gains. The travel managers gaining influence are the ones who understand AI's opportunities and its risks, and can advise their organisations on both.

There's a measurement gap to close, though. Half of buyers say it's too early to judge AI's ROI — but only 21% actually measure it, and those who do overwhelmingly report positive returns. If you want budget and influence, build the measurement framework now.

Looking ahead: agentic AI, unified platforms and MCP

The panellists agreed on the direction of travel. Lempart predicted agentic AI will soon make and execute booking decisions — always with a human in the loop — finally delivering the long-promised sub-three-minute booking. Jagers challenged buyers to push time savings from 70% towards 80–90% within 12 months. Goswami pointed to unified, omnichannel platforms and emerging standards like Model Context Protocol (MCP) — an open standard that lets AI systems connect to many content providers through one layer, rather than hand-crafting every integration — as the plumbing that will make it possible.

The message for travel leaders: the technology is largely ready. The question is whether your data, your governance and your organisation are.

Frequently asked questions

What is AI used for in corporate travel management?
The most common uses are traveller-servicing chatbots (52% of buyers) and automated reporting (44%). The most *effective* uses, according to BTN's 2026 research, are expense auditing and fraud detection, RFP and sourcing support, disruption management, forecasting and supplier performance analysis.

Will AI replace travel managers?
Unlikely in the near term. Only 4% of surveyed buyers are concerned about their long-term relevance, while 16% say AI has made their role more strategic. Travel programmes still need human judgement on policy, governance, supplier strategy and duty of care.

Do companies need a formal AI policy for travel?
Yes — and most don't have one. Only 29% of organisations have a formal AI policy covering travel, despite tightening regulation such as the EU's AI rules formalising in 2026. At minimum, buyers should establish guardrails on data use, require AI schedules in supplier contracts, and audit how supplier models are trained on their data.

Is AI in business travel delivering ROI?
Where it's measured, yes. Only 21% of organisations formally measure AI ROI, but that group reports strong returns. Panel examples include up to 70% less time spent on travel and expense processing, 90%+ compliance improvements, and 20–30% customer service productivity gains.

-