Will agentic AI become your travel agent?

There’s a bit of friction in travel planning for everyone. We all accept the clicks and comparisons, but there’s a frustration in stitching together a journey from disconnected parts (and always knowing you’re probably missing a better option for at least part of your journey).

Flights live in one system, accommodation in another, experiences somewhere else again. Even the most polished interfaces still ask the traveller to do the work of joining the dots.

Over the last decade, travel startups have focused on smoothing that surface. They made search faster and booking seamless; payments collapsed into a single tap. Each step has reduced friction, but the structure underneath is still the same: the traveller is the manager.

McKinsey’s recent work on agentic AI suggests that structure is beginning to change. Instead of tools that support decisions, a new class of systems is emerging that can make and execute those decisions on behalf of the user – even adjusting plans in response to real-world changes. 

It’s a shift from interface to operator.

The moment users hand over control

For founders in the travel tech space, this creates a new kind of opportunity. The question is no longer how to help users choose better, but how and when they are willing to hand over control.

There is a clear appetite for reducing complexity. Travel asks people to navigate fragmented information, shifting prices and unfamiliar rules. The appeal of a system that can manage that complexity is obvious. 

But at the same time, the report makes clear that trust is a major boundary. Travellers are more comfortable with AI suggesting options than committing to them, particularly when payments, high-value bookings or multi-leg journeys are involved.

This tension creates a more nuanced design space than full automation. The most effective products in the near term are likely to feel more like co-pilots than autonomous agents – systems that simplify and guide, while keeping the user visibly in control. In this context, trust is built gradually – through reversibility and consistent performance.

Founders can create value in chaos 

Planning is where this shift begins, but disruption is where its value becomes tangible. Travel is uniquely vulnerable to change. Delays, cancellations and missed connections can unravel even the most carefully constructed itinerary in a matter of minutes – and create a lot of stress. 

McKinsey points to disruption management as one of the most compelling applications for agentic AI. Systems that can:

• Detect issues in real time
• Rebook alternatives
• Replan the journey (without requiring the traveller to start again from scratch)

For a startup, this is a powerful entry point. Planning might attract users, but recovery during stress will create loyalty. A product that can step in at the moment of friction and resolve it for you occupies a very different position in the user’s mind.

The new interface layer

Beneath these potential use cases is a more structural shift around where value sits. Because if travellers begin to rely on AI agents as their primary interface, the traditional front doors of travel (airline sites, hotel brands, even aggregators) recede into the background. The interaction layer moves to whoever owns the agent.

And that raises the stakes. Control of the customer relationship may become less tightly held by the platforms that aggregate supply – new players on the market will have a chance to define this layer from scratch, while incumbents have to face the challenge of remaining visible within it.

The opportunity ahead 

Travel has always been one of the most fragmented consumer industries, and it’s that fragmentation that has created both complexity for users, and inefficiency for providers. Agentic AI reframes it as an opportunity. Where systems are disconnected, orchestration becomes valuable. And where decisions are layered and time-sensitive, delegation becomes meaningful.

So the startups that define this next phase might not look like travel companies in the traditional sense. They’ll look more like operators – persistent, adaptive systems that sit between the traveller and the world, turning a fragmented process into something coherent and (hopefully) easier. 

Discover the future of travel at LEAP from 31 August – 3 September 2026. Meet the founders building new layers of travel, and connect with global travel tech leaders.