AI will not take your job. The agent who uses AI will.

Everyone Asks If AI Will Replace Travel Agents. Wrong Question. Ask Instead: Which Ones?

Something happened a few months ago that I keep thinking about.

A company called Long Lake paid 6.3 billion dollars for AmexGBT. The biggest corporate travel company in the world. A business that is basically thousands of travel agents serving corporate clients.

OK, big acquisition, fine. These happen.

But here is the part that stayed with me after reading the related article at Skift originally. Before Long Lake came in, the bankers running the sale contacted 64 potential buyers. 46 of them said no. Most of them did not even sign the confidentiality agreement to look at the numbers. And the number one reason they gave? AI. They were afraid that a business built on human travel agents does not have a future.

Think about that for a second. These are people with billions to spend. They looked at the American Express name, decades of client relationships, 2.7 billion in revenue... and they walked away.

So when someone asks me... is AI going to kill the travel agent... I take the question seriously. Because apparently 46 investors already answered yes.

My answer, after 28 years in this industry: no. And also yes. Depends what kind of agent we are talking about.

What AI is actually taking

I will try to be concrete because this debate is usually just noise.

AI is good at the routine stuff. Searching availability. Comparing prices. Making a basic itinerary from a destination and a budget. Answering the standard questions... baggage, visas, cancellation rules. Writing the follow-up email. Processing a simple booking.

If that is most of your day... then yes, you have a problem. Not because AI will replace you next month. But because the clients who used to call you for these things are already getting the answers somewhere else. Faster. For free. This started years ago with the booking engines anyway. AI just pressed the accelerator.

Now here is a number worth knowing. Phocuswright did research on why people book with a travel agent. Price? Only around 2% said price. The other 98% come for something else. The relationship. The expertise. Someone who picks up the phone at 11 at night when the flight is cancelled and the hotel says they have no booking.

Can AI do that? No. Not today. And for expensive complicated trips, honestly, not for a long time.

What AI cannot do

I spent a big part of my career with airline commercial structures, cruise, Fly & Cruise, luxury travel, so let me describe what selling these products actually looks like.

A client wants to cruise the Greek islands. Flying from New York. 3 nights in Athens before the cruise. Embark Piraeus, disembark Istanbul, fly home through London with one night stopover. There is a budget. A preferred cabin category. Dietary requirements. Their anniversary falls on day-4 and they want something special on board. Oh, and part of the family is coming from Sydney and everyone needs to land the same day.

Ask AI for cruise options and you get a list in 30 seconds. Nice. But who builds this whole thing so it actually works? Who handles the air and sea coordination? Who knows that one cruise line's embarkation in Piraeus in August is smooth and another one is chaos? Who calls the ship to arrange the anniversary dinner? And who takes the phone call when something breaks?

That is not a booking. That is a professional service. And professional services do not disappear because a chatbot exists.

Actually something else is happening. The more AI floods everyone with generic travel information, the more valuable the person with real specific knowledge becomes. When anyone can get a basic itinerary from a chatbot... the question changes. It becomes: who do I trust to make sure this actually works?

The cruise angle... this is where it gets interesting

I want to stay on cruise for a moment because this is where I see the biggest gap between what AI can do and what the job actually requires.

A cruise is not a flight plus a hotel. It is a bundle. Accommodation, transport, food, entertainment, excursions, port logistics, often land arrangements before and after... all in one purchase. One cruise booking can mean thirty decisions and every one affects the others. Ship, itinerary, cabin, deck, dining, air connection, transfers...

There is a new AI platform, launched this year, built specifically for cruise booking. Gangway Technologies. Their own analysts said something I found very honest. Yes, AI can take away the pain of comparing 5,000 sailing options. But the cruise agent's role, they said, will shift... from search agent to consultant.

Read that again. Even the people building AI for cruise are not saying agents disappear. They are saying the searching disappears. Big difference.

And that matches exactly what I see in the market. The agents doing well right now are not fighting AI. They let AI do the research and the admin. And they spend the saved time on what AI cannot touch... advising, coordinating, being personally responsible when it matters.

Back to the 6.3 billion

So why did Long Lake pay all that money when 46 others ran away?

Same reason. That is the beautiful part. The 46 saw AI as the thing that kills the business. Long Lake saw AI as the thing that transforms it. They did not buy AmexGBT despite the AI risk. They bought it because of it. The plan is to rebuild the biggest travel platform in the world with AI in the middle... not to remove the agents. To make them faster and stronger.

Alex Taubman, Co-Founder & CEO of Long Lake said it himself when the deal was announced: the future of business travel is AI and human agents working together (his words, more or less). The man just paid 6.3 billion for the biggest TMC on the planet. If he believed agents were finished, that is a strange way to spend the money.

46 buyers were afraid of disruption. One buyer bet on transformation. The agents who survive the next years are the ones who understand which side of that bet they want to be on.

So what does the surviving agent look like?

From what I see working right now... 4things.

One. They sell complexity, not information.

Fly & Cruise. Multi-stop luxury trips. Groups with special requirements. Corporate travel where someone's safety is your responsibility. Products where a mistake costs real money and the client needs a human being who answers for the result. Information is free now. Judgement is not. Responsibility is not.

Two. They use AI as a tool, not an enemy.

Let AI do the first search. The first draft itinerary. The follow-up email. Take the hours you save and put them where only you can deliver... the client conversation, the supplier relationship, the fixing when things break. And listen, the agents losing business right now are mostly not losing to AI. They are losing to other agents who started using AI before them and now work three times faster.

Three. They are known for something.

When everyone can get generic advice from a chatbot, being the person who really knows Greek island cruising, or Fly & Cruise from Southern Europe, or airline commercial structures from the inside... that is what clients search for. A niche. A visible reputation. A blog, a LinkedIn presence, whatever. AI cannot copy a reputation.

Four. They know the product with their own eyes.

I read a comment recently that stuck with me. For every agent who really knows cruising, there is another one who pretends. And if an agent recommends you a ship, ask when they last sailed on it. Because a cruise from15 years ago tells you nothing about the product today. AI can list ships all day. It cannot tell you which deck to avoid on a specific vessel because of the anchor noise at 06:00 in the morning. That comes from being there. And being there does not scale.

The uncomfortable part

I will say this directly because I think the trade press is too polite about it.

There are agents out there whose whole business is things AI already does better. Searching flights the client could find alone in 30 seconds. Booking hotels from a database the client can access directly. Building itineraries by copy-paste from supplier websites. And charging a fee for it.

For them, the disruption is not coming. It is here. And not because AI is some miracle. The value was thin already... the market is just discovering it faster now.

But for the agent who knows the product, keeps real relationships, works in complex travel, and treats AI as a productivity tool... honestly, the opportunity is bigger than it has been in years. All this AI-generated sameness everywhere makes genuine expertise easier to spot, not harder.

Where I stand

I have been in this industry long enough to have heard the funeral of the travel agent announced many times. When the CRS came. When the GDS came. When the internet came. Online booking engines. Mobile. NDC. Every time, the same story... the agent is finished. And every time, the agents who adapted came out stronger than before.

Is AI bigger than all of those? Yes. I will not pretend otherwise. The speed of this one is different.

But the question a traveller asks before spending serious money on a complicated trip... that question has not changed in fifty years. Who can I trust to make this work?

AI cannot answer that. A good agent can.

So, to be direct:

•      If you sell complexity and know your product with your own eyes - AI is the best assistant you ever had. Use it.

•      If your job is mostly searching the GDS and copy-pasting itineraries - the change is already here. Move now.

•      If you run commercial for an airline or a cruise line - put your money on the agents who specialise. They are the ones who will still be selling for you in ten years.

•      And if you are an investor... Long Lake just showed you the answer. AI plus the human. Not AI instead of the human.

The travel agent is not finished. The travel agent who refuses to change... that is another story.