Consumers are already taking the advantages of AI. They are moving to natural language search and delegating their requests to agents that research on their behalf. Trust is not yet established, but it is growing, and as it builds alongside spend and itinerary complexity, there is an opening for technology operators to provide the guardrails.
At Phocuswright Europe there was a common theme about who is responsible when an agent gets things wrong. For me it depends on whether the agent is self-installed or offered as a service, much as a tour operator or travel agent is today. Trust may be solved by insurance or by policy. It is too early to say, but it is best earned through explicit actions rather than autonomous ones.
We worried about disintermediation last year, and still do, but less so. The view was that the big four will endure and evolve, and niche players will hold where they own first-party data or an exclusivity advantage. What is certain is that we all need to place bets. AI now makes experimentation affordable, both client facing and internal. Do nothing, lack data, and a startup with a handful of people amplified by AI could take your business.
The situation
Adoption is racing ahead of trust
People are adopting AI for travel at the fastest rate the industry has measured. In the United States, the share of travellers who used AI for travel rose from 33% to 56% across 2024 to 2026, and the European markets roughly tripled. Travel remains one of the things people most want to spend on. The capability is compounding. What follows is the complication.
Trust lags capability: about a quarter of European travellers would let AI book. Supply is exposed too: a frontier model was pulled from partners under export controls, so a government, not just a vendor, can cut your access. Distribution is fragmenting.
So the question for a leadership team is narrow and urgent. Where should you invest over the next two to three years, and how do you avoid being commoditised by an agent or an AI-amplified startup? The answer is to win the AI shop window while trust still lags, build the data and architecture that cannot be copied, and reset the commercial model for an agent-led market. Three moves carry it.
Move one
Win the AI shop window
AI is the front door, and trust, not capability, sets the pace. Discovery is moving across three webs that now sit side by side: the human web where people navigate site to site, the LLM web where an assistant helps you choose, and the agentic web where software acts on your behalf. Visibility starts with strong SEO, because the models answer by searching the live web. But agents read information, not pictures, and most sites still block the bots that agents use. Brand is becoming reviews, reputation and reliability.
Stand up a generative engine optimisation programme, make your inventory machine readable, and let genuine agents through. Build trust the way people grant it, with visible options, shown savings and a human approval step, and automate the low-value, repeat, structured trips first. Corporate travel is the near-term win, not the family summer holiday.
Move two
Build the data and architecture moat
AI is only as good as the data in its memory, so poor data in means poor results out. The lasting assets are a memory of the product and a memory of the guest, plus the first-party data that the middlemen cannot take. Providers will keep changing, and as the export-control episode showed, the change is not always commercial. That is the case for a switchable architecture: a thin abstraction layer with portable prompts and evaluations, so swapping a model or a provider is a routing change rather than a rebuild or a crisis.
The economics reward doing more, not cutting cost. Inference gets roughly ten times cheaper each year while labour stays flat, and AI-first firms earn many times more revenue per employee. Fund data hygiene as a programme in its own right, build or buy one joined-up memory layer, own your first-party data, and put a switching layer between your products and the models. Use AI to scale output through your people, not to shed them.
Move three
Reset the commercial model
Loyalty is moving from points to recognition, personalisation and recovery when things go wrong. Payments are getting agent-ready, with capped virtual cards and programmable authorisation, which is the part that matters for agentic commerce; smarter FX helps, but it is operational, not the core prize. Distribution is fragmenting as banks, cards and super-apps start selling travel on travel infrastructure, and much of the growth, and the defence, sits in B2B, where the moat is supply, the data you own and execution. Decide on purpose whether you are the front door or the infrastructure, and invest behind that choice.
Where to play
The same shift, different priorities
The thesis holds across the industry, but the first move depends on where you sit.
Why this holds
A strategy built for uncertainty
The bear case is fair. Agentic booking may stay niche for years if trust does not move. The incumbents may simply reinvent. Access, certification and inconsistent versioning may hold the industry back regardless of model quality, and data security, the single biggest stated barrier, may slow adoption inside large firms.
The design answers all of it. The no-regret moves, clean data, first-party data, visibility, agent-ready payments, literacy and a switching layer, improve the business today and cost little in flexibility. You have not bet the firm on one future. If the fast scenario arrives, you scale the contingent bets. If the slow scenario holds, you have run a tighter, better-instrumented business. You win either way, which is the test of a strategy built for uncertainty.
It really is game on
Those who bet and experiment will stay ahead, and the gap for the laggards grows wider each week. DataArt partners with travel and hospitality businesses to turn this into working systems: data and memory foundations, a model-agnostic AI architecture, AI-ready datasets, and agent-ready visibility and payments. A short diagnostic on data readiness is the natural place to start.
