Both in the room.

At Travelopia, we've stopped asking whether AI can help. The more interesting question is what kind of help it actually is.

Because the moment you look closely at where work actually gets stuck, you realise the friction has two parents. There's the technology side — the tools that don't talk to each other, the manual steps that shouldn't exist, the data that lives in the wrong place. And then there's the people side — the unclear ownership, the habits that formed around broken systems years ago, the team structures that made sense once but don't anymore.

Fix only one and you've done half a job.

The tech solution that doesn't land

I've seen well-designed AI solutions fail to stick at Travelopia because the organisation around them wasn't ready. A workflow gets automated, and it's technically impressive — faster, cleaner, more consistent. But the people whose jobs just changed weren't involved in designing it. Nobody thought about what they'd do with the time they'd saved, or whether they trusted the output, or how they'd know when something had gone wrong.

So they work around it. Or they use it inconsistently. Or they wait to be told whether they're supposed to like it.

The tool worked. The change didn't.

The change programme that doesn't deliver

The opposite also happens. Invest heavily in the human side — change management, communications, training programmes, workshops on AI mindset — and you can produce a lot of enthusiasm with limited results. People feel informed and included. They understand why AI matters. But the actual tools are clunky, the data is unreliable, and the integration into daily work is shallow.

Engagement without capability goes nowhere.

Why you need both in the room

Sree and I talk about this constantly in terms of left hand and right hand. The technology decisions and the organisational decisions have to move together — not take turns. When one hand doesn't know what the other is doing, you get well-designed tools that land badly, or well-intentioned change programmes that have nothing solid underneath them.

At Travelopia, this has become one of the clearest lessons. The tech side needs a seat at the business table — and right now, that matters more than it ever has. Not as a vendor of solutions, but as a genuine partner in understanding where the organisation is trying to go and what's getting in the way. When technology turns up late to that conversation, it spends the rest of the time catching up.

And here's something that gets lost in the current excitement around AI: the answer isn't always AI. Sometimes the right fix is a solid piece of technology with no AI in it at all. Sometimes it's automation — clean, reliable, deterministic. Sometimes it's AI-enabled, where the intelligence accelerates or enhances something that still needs a human in the loop. Sometimes it's AI at the core. Getting that call right matters enormously — for cost, for adoption, for outcomes. But you can only make it if tech is in the room early enough to ask the question. That's exactly why the seat at the table isn't just about delivery — it's about diagnosis.

But the other half of this is just as important — and it's the part that's easiest to underestimate. Senior leadership sponsorship isn't optional. If the people at the top aren't visibly engaged, aren't asking questions, aren't making it clear this matters, the organisation reads it as optional. The initiatives that have moved fastest at Travelopia are the ones where leadership didn't just endorse from a distance, but showed up.

And yet sponsorship alone doesn't make it real. The people who actually do the work — the real users, the ones whose daily routines you're asking to change — have to be part of building it. Not consulted at the end. Not trained on something finished. In it from the start. They know where the process actually breaks. They know what the workarounds are. They know what will get used and what won't.

You can't do it without senior buy-in. You can't do it without real users either. And you definitely can't do it with the left hand and right hand working in separate rooms.

What this looks like

It means bringing both lenses to the same problem at the same time. When you're mapping a workflow that AI could improve, you're asking simultaneously: what does the technology make possible, and what does the organisation need to change to make that real? Not sequentially — together.

It means the technical design reflects how people actually work, not just how the process is supposed to work on paper. And it means the change approach is specific enough to address the actual friction, not generic enough to apply to anything.

It's harder to run. It requires different people to share a room and speak each other's language. But it's also the only version that consistently works.

Where the friction lives

The friction is almost never purely technical or purely human. It's the junction between them — the point where a capable tool meets an unprepared team, or where a willing organisation is let down by tooling that doesn't hold up in practice.

That junction is where the real work is. And it's where most AI programmes run out of road.

The next posts in this series will go deeper on both sides — what good looks like on the technology front, and what good looks like on the people and organisational front. Because understanding them separately is useful. But knowing how they interact is what actually moves things forward.