Travel management companies are expensive businesses to run. They need consultants, support teams, account managers, implementation teams and reporting layers around the technology. The more valuable the client, the more complex the service often becomes. Government travel, global corporates, consulting firms and senior executive travel are rarely clean, fully digital workflows.

The last wave of TMC disruptors attacked the competition with software. Navan, TravelPerk, Serko and others made corporate travel look more scalable: better interfaces, more automation, tighter expense integration, faster support and less dependence on traditional manual service. AI now gives the whole industry a chance to do that again, but at a larger scale.

For established global TMCs, this is the opportunity. If it works, scale becomes an advantage again because large TMCs have more workflows and more service volume to automate.

But the same technology is also the threat. As AI makes workflows easier to automate, the barriers around the TMC business weaken. Startups can build faster. Expense platforms can move deeper into travel. Corporate card providers can add trip workflows. ERP systems can extend into travel control. Large travel platforms with inventory and service operations can package cheaper corporate travel tools.

That is the risk investors are now trying to price. In 2026, AI is no longer just a future product theme for TMCs. It is showing up in M&A decisions, credit facilities, margin claims, enterprise integrations and platform roadmaps. The clearest signal came from American Express Global Business Travel. During the company’s sale process, Skift reported that 46 of 64 potential buyers declined to pursue Amex GBT, with AI disruption risk cited as the top reason. For some investors, AI made the world’s largest corporate travel platform look vulnerable.

Then Long Lake agreed to buy Amex GBT for $6.3 billion. Why?

Amex GBT is the main test case because it sits at the centre of both arguments.

The company completed its CWT acquisition in 2025, adding more scale to a group that already included Egencia and Ovation. That scale is complex to manage, but it also gives Amex GBT a large base of service work that AI can potentially improve.

For a large TMC, AI is about changing how service work gets done, as human agent becomes more productive, empowered by AI.

That is the Long Lake thesis. The buyer is not treating Amex GBT as a legacy agency to protect from AI. It is seeing it as a large operating platform where AI can improve booking, disruption management and travel administration.

This is why the transaction matters beyond one company. It shows how investors are starting to separate TMCs into two groups: those exposed to AI automation, and those capable of using AI automation to improve margins.

The next issue is fragmentation.

A business trip does not start and end with a booking. It moves through policy, approval, payment, support, receipts, expense categories, cost centres and reporting. Historically, those steps often sat across separate systems. The traveller booked in one tool. Payment happened somewhere else. Finance reviewed the receipt later. Travel managers checked compliance after the fact.

AI makes that harder to defend. An AI system cannot act well if the context is split. It needs to know the traveller, the policy, the payment method and the expense logic. Otherwise it can answer a question, but it cannot reliably complete the workflow.

This is why SAP Concur and Amex GBT’s Complete matters. Complete is being developed as an AI-enabled solution for booking, servicing, payments and expensing. SAP’s Joule copilot is being layered into the experience, with automated help and handoff to a live Amex GBT travel counsellor when needed.

Navan and Perk are moving toward the same problem from different starting points. Navan began with a software-native travel-and-expense model. Perk, formerly TravelPerk, has moved from modern business travel into broader travel and spend management. The target is the same: less administrative work around the whole trip.

Perk’s 2026 financing shows why this matters commercially. The company secured a $300 million credit facility to accelerate growth of its AI-native travel and spend platform. More interesting was the margin claim: Perk said AI deployment helped move gross margins from around 40% to the mid-70s over three years.

That is the kind of evidence the market will increasingly demand. AI announcements are easy. Margin movement is harder.

But there is a second-order question. If AI makes TMCs materially more efficient, corporate clients will eventually ask for some of that benefit. Buyers have always pushed on transaction fees, service levels and account support. If suppliers can show higher margins from automation, large clients may ask whether pricing should come down, whether service should improve, or whether savings should be reinvested into better reporting, duty of care and traveller experience.

That could make AI both a margin lever and a pricing weapon.

This stretches the category. The old boundary was business travel booking and support. The new boundary is controlled movement of people and spend.

The next pressure is where the traveller starts.

Navan’s launch of Navan Anywhere is useful here. The company is putting its AI travel agents inside Gemini Enterprise. The employee can start inside Google’s enterprise AI environment, while Navan provides the travel inventory, policy logic, booking capability and expense automation underneath.

This matters because the TMC interface was already becoming less sacred. Corporate clients have long treated TMCs as partly interchangeable. They negotiate fees, switch providers, compare service models and ask what the next TMC can do better or cheaper.

AI adds another layer to that pressure. If travellers can start in Gemini Enterprise, Microsoft 365, Slack, Teams, Concur, an expense tool, a calendar or a company portal, the TMC does not win by forcing every interaction back into its own app. It wins by being present in more of the places where work already happens.

This is different from the OTA problem. For leisure OTAs, losing the consumer interface threatens the brand relationship and the booking path. For TMCs, the interface was never the whole business. Corporate buyers already expect policy, duty of care, payment, reporting and servicing as standard.

So the question is whether TMCs can stay commercially relevant when its service appears through someone else’s interface?

Corporate travel cannot use open-ended AI in the same way leisure travel can. A traveller may ask for the cheapest flight, but the cheapest flight may break policy or create operational risk. A hotel may be in budget but outside the approved area. A visa answer may depend on nationality and purpose of travel. A disruption may require judgement, not only rebooking.

This is where FCM’s approach is relevant. The company is relaunching Sam as an AI assistant across more than 90 countries, but the important point is the positioning: AI inside guardrails that respect travel policy, traveller profiles, spending thresholds, approval workflows and supplier preferences.

BCD is making a similar governance argument. Its 2026 AI messaging warns that the industry is using “AI” as a catch-all term for very different systems, from basic automation to tools that can act more independently.

That distinction matters because corporate buyers need to understand effort, ownership, liability and control. A global company cannot let AI agents make uncontrolled travel decisions because the interface feels smooth. It needs audit trails, data security, explainable policy logic and human escalation.

This gives established TMCs a defensive advantage, but only if they move fast enough. Their pitch cannot simply be “we have AI.” It has to be: “we can make AI safe enough for managed travel.”

AI will not disrupt the TMC industry in one simple direction. It will make some TMCs look more exposed and others more valuable.

For TMCs, the category is expanding and narrowing at the same time. Expanding because travel, expense, payments and workflow automation are converging. Narrowing because clients will tolerate less manual work, less fragmentation and less opaque pricing.

The next TMC leader will not be the company that adds AI to travel management. It will be the company that makes managed travel cheaper to run, easier to control and easier to plug into the systems where work already happens.