Why mergers and AI transformations don't mix — the hidden cost inside the GBT deal

Three companies. Three teams that had spent years competing against each other. Then one day we were told we were one company now.

I was in that room. When Hotelbeds, GTA, and Tourico became one organization, I led the technical front line. So when Long Lake paid $6.3B for American Express Global Business Travel last week, I didn’t read the headline the way most of the industry did. The price isn’t the risky part. The timing is.

GBT is being asked to do two enormous things at the same time, on the same organization, competing for the same scarce resource. And that resource isn’t money. It’s leadership attention and the willingness of people to take risks.

The bet nobody is pricing: trust

Here’s what the integration math leaves out. GBT went from 19,000 to 27,000 people overnight after pulling in CWT. The deal carries a $155M synergy target, and “synergy” is the word you use when you mean headcount.

I have lived the inside of that number. When you merge teams that competed for years, everyone in the building knows that three talented people are now doing every job, and two of them will be gone. Nobody says it out loud. But it shapes every conversation. Every coalition. Every decision that gets made a little too slowly because the person who could make it is busy protecting their position.

The hardest part of the merger I ran was never the systems or the strategy. It was trust. Different cultures, different survival instincts, thrown together at the exact moment we needed to be at our absolute best for the merger to work at all. That problem is already live inside GBT today, and it’s the part that doesn’t show up in any investor deck.

Now stack a second transformation on top

That’s only the first bet. Long Lake is a firm built around AI-modernizing the businesses it buys. So the plan isn’t only to integrate CWT. It’s to run an AI transformation on the merged entity, at the same time.

Not a pilot. Not a chatbot bolted onto the support desk. A structural rebuild of how the work actually gets done. That kind of change requires the hardest organizational decisions a company can make, and most travel companies are still delaying those decisions even without a merger running in parallel.

Think about what that asks of one employee. Survive the merger cut. Then survive the AI cut. And you don’t even know whether one of the two people who survive the first round makes it through the second. That is not the environment where anyone makes a bold move. People in survival mode optimize for staying invisible, not for rebuilding the company. A $6.3B thesis needs speed. Fear produces the opposite.

The companies carrying none of this weight

While all of that plays out inside GBT, the competition is unburdened.

Navan and Perk, the company formerly called TravelPerk, were AI-native from day one. No legacy stack to unwind. No merger hangover. No two cultures to reconcile before anyone can ship. They raised capital specifically to build for this moment, and they don’t have to convince 27,000 anxious people to change all at once. They just build.

That’s the real gap. It’s not a feature gap or a funding gap. It’s a friction gap. One side spends its energy managing internal uncertainty. The other side spends all of it moving forward.

GBT is the extreme version of a problem everyone has

I want to be careful here, because it would be easy to read this as a story about one big deal. It isn’t.

Every legacy travel company is running this same race right now. Transform to AI while keeping the business you already built alive and paying the bills. GBT is just the loudest, most expensive example. The difference for most companies is that they don’t have $6.3B and 27,000 people to absorb the friction. They have a smaller margin for error and a shorter runway to figure it out.

I’ve been inside this exact problem, and I know what it costs when transformation has to fight its way up through layers of organizational fear. That cost is invisible on a spreadsheet and enormous in reality. It’s the reason I’m building Bitravel the way I am: AI-native, no legacy to defend, the agent — we call it Alex — built into the product from the first line of code rather than retrofitted into an org that’s also trying to survive a reorg.

The window to make this transition is smaller than the press cycle suggests. The question I’d ask any travel leader reading this is simple. Where is your organization in this transition, and what is actually slowing you down — the technology, or the fear in the building?

This is the kind of problem we work through building Bitravel — an AI-native business-travel platform built from scratch, without a legacy org to drag along. Book a 30-minute call for a live look, no deck.