Why TMCs must embrace agentic AI now

Justin Schuster

Welcome to the Travel Is a Human Emotion podcast. My name is Justin Schuster. I’m the SVP of Marketing at Spotnana. I am extremely excited and delighted to share that our guest today is Martijn van der Voort. Martijn is an independent travel industry advisor and the founder of AstraNomad, a consultancy focused on the intersection of travel technology and agentic AI.

We’ll probably get into this, but Martijn has a very strong opinion on what exactly defines agentic AI. That opinion is shaped by spending nearly 16 years at CWT, where he held senior roles in product delivery, giving him deep operational insight into how traditional TMCs are built and where their limitations sit today. Among other responsibilities, he serves as a non-executive director at Murphy AI, a company focused on building deterministic governance infrastructure for agentic commerce. Martijn is widely known for his candid and often challenging perspective on the future of travel, particularly around AI, distribution, and the shift from legacy models to agentic execution.

Among the very valuable roles that Martijn plays in the travel industry, and among the reasons I’m so excited to have him on our show, is that he is able to say a lot of the things that people in other positions may not be able to speak as freely about. I think today will be a very interesting conversation for that reason among many. Martijn, maybe we can kick things off by asking you — given your deep knowledge of TMCs and how they operate, and your experience at CWT over so many years — what do you think are the key challenges that TMCs are wrestling with today?

Martijn van der Voort

That’s a very good question, Justin, and a nice kick-off. TMCs are full of very clever people. I don’t think the real issue is that TMCs lack ideas. The issue seems more that they’re still carrying an architecture that was built for another age than the one we’re now in.

If I take a step back — and the reason I spend a lot of time evangelizing and trying to change the industry — it really is about a sense of purpose. I truly believe that the industry can move faster and deliver better outcomes for everyone: for themselves, and for their clients.

Being an independent voice does allow me to be a bit more direct, as you alluded to. A lot of what I say publicly reflects conversations happening privately with me across the ecosystem. People tend to recognize the same constraints, but they’re definitely not in a position to articulate them, very often because of commercial ties. And I also don’t share my views just to be proven right or wrong. I share mainly to expose people to a different perspective, to encourage thinking beyond the current narratives and the echo chambers. We need to keep talking. We need to drive conversations forward.

Justin Schuster

I couldn’t agree more. I don’t know how else we move the industry forward. I appreciate the role you play and the expertise you bring to the conversation happening on LinkedIn and in other places. One of the things you are very vocal and knowledgeable about is AI. It seems unavoidable right now — it’s in every news cycle, and there’s always something new to be said about the topic. I’m curious to get your take: what do you see as the biggest opportunities around AI in corporate travel that you feel are achievable in the next 12 months or so?

Martijn van der Voort

Near-term AI opportunities are real. They’re more operational than magical. A lot of noise sits at the extremes — either everything is about to change overnight, or nothing meaningful is happening yet. The reality is a lot more practical.

What is achievable in the near term? The first area is service efficiency. Especially in corporate travel, it’s great that we can book with agentic AI, but if we can’t service anything properly afterward, it’s not going to work for the traveler — and that’s what we need to keep in mind. The opportunities are in disruption handling, itinerary changes, email parsing, and policy checking. A lot of what used to happen in very heavy mid-office workflows — PNRs rolling through and all of that — AI is already massively improving the speed and reducing the friction there.

The second area is decision support: helping travelers and travel managers navigate the many choices they have around content, applying policy and context, and doing research in that area as well.

And then there’s the whole conversational experience. Everyone is moving toward conversational AI — what some call chatbots. People are increasingly expecting to interact through chat, voice, and maybe in the future other wearables. Could be glasses. You see a flight you want to take, and it’s handled just like that. It could well be.

Justin Schuster

For sure. As I’m hearing you describe the landscape, I agree with every part of it. AI really has applications that are pervasive — everything from the planning experience and how I interface with a system that can book for me, all the way through to servicing, analytics, and how I understand the health of my program. When you think of it in such a pervasive way, what do you think needs to be in place from an infrastructure perspective to truly take advantage of AI?

Martijn van der Voort

That’s a very solid question, because when we’re going to start trusting something with full autonomy, technology needs to earn that trust first. We’re talking about areas like authorization, liability, and payment control. That requires a certain architecture, and that’s not easy to achieve. You cannot let agentic systems run without deterministic governance. It’s not just about a better interface — yes, you want to move away from online booking tools as they are, form-based and web-based. But what it’s truly going to be about is whether a system can transact and act safely on your behalf. We don’t want an agent making a decision that doesn’t comply.

It really comes down to three layers. The first is governance: what is the system allowed to do? The second is execution: can it actually complete the task — book a flight, make a transaction, whatever else? And the third is audit: traceability, because we always need to understand what actually happened. Who initiated the action? Was it the agent? Was it you? Did something go wrong? Without those layers, you have assistive AI. You don’t have true agentic capability at all.

Justin Schuster

That absolutely makes sense. I appreciate the point about needing an audit trail — both for the system to learn and to troubleshoot any issues that emerge. That’s a piece people may still be thinking through in terms of implementation and execution. It begs the question that’s top of mind in the industry right now: will the LLM providers become agentic systems and actually be able to book on your behalf? Or will they focus on discovery and trip planning? I’m curious to get your thoughts on how you think things might shape up over time.

Martijn van der Voort

We’ve seen OpenAI do a U-turn on their pay-within-the-app model. That whole idea of an app within an app that certain big players have been exploring — that’s not agentic AI. You still have that layer in between. I’m not so sure that the large LLMs are even well suited to do this, for a variety of reasons. Travel may not even be their priority. They’ve got bigger fish to fry in terms of how they’re shaping their overall companies.

My guess is Google has a chance, because theoretically they’ve got a lot of different things that can be drawn together to become a player. On the other hand, you could argue: why have we seen so little movement from such a powerful company? I’m not so sure they’re as well positioned to actually achieve it.

Whereas there are fringe players and edge players in far more interesting fields at the moment — fintech, other technology companies that don’t necessarily approach agentic AI from a travel angle. They just see travel as another node in a much wider universe of nodes.

Justin Schuster

I agree. The more I understand about travel, the less I feel I have a handle on its complexity. It seems infinitely deep, and building true automation around it takes years of work to handle all the corner cases and special circumstances travel requires. It’ll be some time before those players are interested in building that on their own. It’s more likely they’ll end up forming partnerships in one way or another to address that need.

Martijn van der Voort

That’s my thinking as well. And also — should we really rely on the LLMs? Should you heavily rely on building platforms around large third-party infrastructures? Not necessarily. I think anyone serious about agentic AI who is going to run agentic AI should own their own LLM, or SLM — multiple SLMs, even.

Justin Schuster

When I think about the evolution happening in parallel — and this is part of what makes it so hard to just plug in if you’re an organization owning their own LLM — the airlines are going through a lot of change right now. Hotels are going through a lot of change. There’s a move to implement new offer and order systems, and to upgrade to new CRS and PMS systems. How do you think this is going to change travel? What do you think the new retailing experience that emerges from these investments will look like? And what does the rest of the ecosystem need to do to get ready to consume the new capabilities that will be available in the market?

Martijn van der Voort

Modern retailing only matters if the customer actually feels a benefit. We’ve had the NDC story, and the questions around it have been debatable. There’s a lot of investment happening around airlines and hotels, and the outcomes need to be tangible.

We’re moving toward more dynamic and contextual retailing — but are we truly moving to less rigid structures, more flexible based on what the traveler wants and needs, their context and needs? For TMCs to fit into the story that airlines and hotels are already building, the TMC role and model needs to change. The role becomes one of orchestration.

If we can bring that complexity together in a way that feels coherent — and obviously following policies, which is very important in corporate travel — and do it consistently, then we can get somewhere.

I feel there are certain TMCs that are developing the right ideas, and I’m not going to name names. But a lot of them don’t necessarily have a grasp of the complexity. TMCs will exist. They will have a future. It just won’t look like the commercial and operational model they have today, because they need to adapt to the needs. If I’m using something in my private life and I’m starting to get heavily used to agentic AI, I want that translated when I’m at work. I don’t want to look at some outdated platform where I still need to use dropdowns and type in my departure and arrival airport, then manually add a hotel on top of it.

Corporate travel — and especially TMCs — actually have a fantastic opportunity, because agentic AI is all about governance, rules, determination, and agency. That’s exactly what we already see in corporate travel today. The customer will never judge by the progress of the architecture. Most of them don’t care about that. They look at outcomes. The traveler will judge by whether the experience becomes easy, clear, and relevant to their personal needs. Translating that orchestration into that experience — that’s where TMCs are going to need to win.

Justin Schuster

That makes a huge amount of sense. You talked about new economics that may evolve over time. I can see how that’s possible given that AI and automation are going to reduce the need for agents to handle transactions — though I don’t think that ever goes away entirely. As that change takes place, how do you think economics might shift in the travel industry?

Martijn van der Voort

In the next five years, we’ll see a bunch of totally different players emerge, and that’s where the economics are going to change. We’re already seeing it today. Big accounts are moving out of incumbent TMCs onto newer platforms, and that’s only going to accelerate. And there are very clever people working on the fringes.

There are several forces at work at the moment gearing up to drive change. The push can come from airlines, from corporates, from fintech, from technology players that most people don’t even have in their peripheral vision.

Economic models will change, for sure. The old transaction-based model, with people sitting in call centers picking up the phone against exorbitant pricing — that’s going to be a thing of the past.

Justin Schuster

One of the other things I wonder might evolve alongside this is how the buyer interfaces with the TMC in a relationship built around trust. It feels as though trust has somehow eroded over the past few years in that relationship, and buyers are increasingly vocal about this. There have been more articles recently about proposals for how economics could evolve in ways that improve trust, or other actions that can be taken. What’s your take on the key reason why buyers feel that trust has eroded, and what do you think TMCs can do to rebuild it?

Martijn van der Voort

One of the things we’ve always heard about is data fragmentation — over the years, TMCs have acquired other TMCs, data gets fragmented, and when you ask for a report, that’s just one example of many issues that result.

On technology innovation, we’ve heard it all: “We’re so innovative for doing this, for doing that.” But what has the actual outcome been for the travel area? Nothing has really shifted. Trust will improve — and I want to choose my words carefully here — trust will improve at the moment that buyers can clearly see how the system actually operates.

Buyers are already asking sharper questions around cost, content, data, and decision making. That’s a healthy thing. But to truly innovate for your own travel program, there are a few paths. You can stick with what you know, because it’s working today and maybe your program doesn’t allow for drastic change. Or you can start educating yourself about what’s around the fringes — look at other industries, see what they’re doing, and translate that back into your own TMC conversations. Start asking the real questions and truly understand what you actually want for your program. We all know the stories about preferred supplier agreements hanging in the background. But is that necessarily going to be right for you as the buyer?

The level of transparency between the TMC, the corporate travel manager, and the buyer is going to become more and more important. The technologies we’re seeing emerge today come fast and hard, and they are confusing and difficult to understand. For trust to be created, there needs to be full clarity. What are we actually paying for? Are we paying for your headcount, or are we paying for innovation on your side? And if we’re paying for innovation, what value is actually being delivered? And where are the constraints? What can’t you do? That transparency needs to be there. What is it that you cannot do? Because that might influence my decision.

Justin Schuster

That makes a lot of sense. I think buyers are looking for TMCs who are willing to be a genuine partner and align with the needs they’re trying to solve for.

Martijn van der Voort

You’re spot on. Transparency is not a risk — it’s actually a differentiator. True transparency means coming together as genuine partners, as equals. You’re going into this new adventure together. That’s where trust and transparency sit. That’s how you create long-term relationships.

Justin Schuster

So as you think about the buyer — and let me ask this question first — the pace of change seems to be only accelerating. One of the new challenges emerging for travel managers over the past couple of years is navigating this period from an innovation perspective. How do I think about all the innovation that’s out there and manage consuming it in a way that’s solving for the highest and best needs of my program? What are your thoughts on how travel managers can approach this?

Martijn van der Voort

I read a lot. That’s the only thing I can say. And I know travel managers don’t necessarily have the luxury of time.

Read a lot, but read subject matter from the right people. And those people might well be outside the travel industry. If we’re talking agentic AI, I probably wouldn’t ask my TMC about it today. I would go to someone heavily embedded in agentic AI, and I would read papers from Harvard or Cornell. It sounds dull, and it’s definitely not bedtime reading. But you need to sharpen up, because the echo chambers are not where you want to get your knowledge. Once you have the knowledge, you have the power — because you can start asking the right questions.

Justin Schuster

That’s a great segue to the next question I wanted to ask. From your perspective — having worked in a TMC for so long and focusing on this intersection of travel, technology, and AI — for the buyers out there thinking about making a move: how should they approach kicking off an RFP for travel this year? There was even a great article in BTN just this week about different approaches businesses are taking.

Martijn van der Voort

I saw that. BTN does a great job of keeping everyone slightly honest. Again: transparency.

I’ll be slightly controversial here, and it may not be liked by many. An RFP should never reward the best narrative. It should reveal the strongest operating model. Start with outcomes. What is it that you want? Be clear on what matters most to you. And then follow up by asking for evidence only. That might be an awkward conversation, but it needs to be a hard one.

If you have partners willing to step into the trust-and-transparency piece with you for the long term, they will have no issue providing the evidence. I’m not talking about just roadmaps. I’m talking about how things are actually working — content, automation, governance, commercials.

We have this multi-trillion-dollar market of corporate travel — anywhere between $2.2 and $2.7 trillion by 2030, depending on who you’re talking to. It’s a massive market, and obviously everyone can have a slice of the pie. The true question really becomes: what share of that value are you designing your program to capture?

An RFP should definitely not test your current capability. It needs to test the readiness for what’s next. What is the future going to look like? That’s what you should be looking at.

Justin Schuster

That’s a great piece of advice. Whatever moment we’re in today, there’s certainly a lot of change happening. Asking your TMC — or the TMCs you’re evaluating — to share their perspective on how they’re preparing for the future, how they’re managing this period of change, and what they need to be ready for tomorrow: I think that’s a great piece of advice.

Martijn, this is the Travel Is a Human Emotion podcast, and we like to ask all of our guests the same question to conclude the show. When you reflect on the connection between travel and human emotions, what comes up for you?

Martijn van der Voort

There is always a human layer in business travel, in corporate travel. There is a plethora of emotions at play. There is stress, ambition, a lot of uncertainty, high energy, and sometimes excitement.

Travel is largely managed by systems, but it’s always experienced by us — human beings. Because at the end of the day, that’s why we do business with each other.

Technology should truly help us as human beings. There are certain emotions we can work to avoid in the future, like stress. Technology should remove friction, not add to it. And once that starts working well, technology fades into the background and allows the traveler to truly focus on the actual purpose of the trip. The purpose of the trip is not going from A to B. The purpose of the trip is to meet and visit someone.

I don’t think people experience travel as transactions. I think they experience it more as a part of life. And I think that should remain at the center.

Justin Schuster

I love that thought. It’s a beautiful way to end. I can just add my reflection that this conversation really started at BTSE almost a year ago, where I had the pleasure of meeting you in person for the first time. I look forward to reconnecting with you in person later this year. Thank you so much for being a guest on the podcast, Martijn. It’s been a pleasure spending time with you.

Martijn van der Voort

Thanks for having me, Justin. My pleasure.