Gee Mann is the founder and CEO of TravlrID
In the past 12 months, virtually every significant platform in travel has announced an AI agent. Airlines are building conversational booking assistants. OTAs are shipping personalisation engines trained on billions of search signals. Hotel chains are deploying AI concierges. Corporate travel platforms are rolling out autonomous agents that can search, book, rebook, expense and report – all in natural language, all without human intervention.
The investment is real. The ambition is genuine. And almost all of it is built on a data foundation so broken that the agents will systematically underperform, frustrate the travellers they were designed to help, and quietly recreate – at far greater scale and with far higher stakes – the exact fragmentation problem the industry has been trying to solve for 30 years.
The travel industry does not have an AI problem. It has a memory problem. And nobody is talking about it.
_____________________________________________________________________
This is a two-part opinion piece by Gee Mann, the founder and CEO of Travlr ID, a travel profile infrastructure company. Read part one here. The continuation in part two below contains what some readers might consider commercial language, which is disallowed in BTN’s policies for opinion contributions. The BTN editorial team has cut a considerable portion of this section, determining that the remainder is illustrative and the topic is of critical importance to the industry.
______________________________________________________________________
Infrastructure always wins
Every major technology shift in travel has eventually produced a layer of shared infrastructure that the whole ecosystem runs on. The GDS solved distribution fragmentation by creating a neutral switching layer between airlines and travel agents. Payment networks created interoperable transaction infrastructure across a global ecosystem of banks, merchants and consumers. Neither of these infrastructure layers was inevitable. Both emerged from a specific moment when the industry faced a coordination problem too large and too expensive for any individual player to solve alone – and when someone built the plumbing instead of another competing product on top of the broken pipes.
We are at that moment again. The AI shift in travel is real and the investment is justified. Agent-to-agent commerce – where an AI planning a trip negotiates directly with an airline’s AI to book a seat, or a hotel’s AI to secure a room – is no longer a thought experiment. The platforms building for it are well-funded and moving fast.
What none of the coverage has adequately addressed is what those agents need to function: a trusted, portable, permissioned profile that travels with the traveller across every interaction, on every platform, with every supplier.
Without that layer, agent-to-agent commerce will be conducted on the same broken data it is trying to automate. The AI will negotiate a room at the wrong hotel in the wrong city because it does not have an accurate address for the event the traveller is attending. It will book a fare class the traveller’s corporate contract does not cover because the policy entitlement lives in a system the agent cannot read. It will miss a visa restriction because that information is in a passport the agent has never seen.
The memory layer is the missing infrastructure. Not another booking tool. Not another expense platform. Not another loyalty programme. The layer underneath all of them – the one that knows who the traveller is, what they are entitled to, where they have been, and what they prefer, and that makes that information available, with permission, to any agent or platform that needs it.
What this looks like in practice
At Travlr ID, we believe we have built this layer. The architecture is built on self-sovereign identity principles: The traveller holds the profile, decides what gets shared, with which platforms, for what purpose, and for how long. When they leave a platform or withdraw consent, access is revoked. There is no central repository of traveller data that can be acquired, breached or quietly repurposed for AI training.
We cannot be a honeypot because we are not designed to hold data – we are designed to route it, permissioned and controlled, to whoever has earned the right to see it. Scania Group adopted it.
The same architecture works in leisure. A traveller using the memory layer does not re-enter their passport on the fourth platform. They share a permissioned, current profile with each, withdraw it at will, and know exactly who has access to what. For platforms receiving that profile – whether a corporate booking tool, a leisure OTA, an airline or a hotel chain – the quality of every downstream AI interaction improves immediately, because the agent finally knows who it is talking to.
The regulatory clock is already running
The regulatory environment is not waiting for the industry to self-organise. GDPR established purpose limitation and data minimisation as legal requirements, not best practices.
The EU AI Act is adding transparency and explainability obligations that will directly touch travel’s personalisation and recommendation systems. Data portability obligations are expanding across jurisdictions.
The opt-in-by-default AI training practices that several platforms are currently running will become legally untenable in key markets within a foreseeable timeframe.
More importantly, the travellers themselves are changing. The generation of travellers entering peak business travel years grew up with app permissions, data dashboards and the right to delete. They are not going to accept “you irrevocably grant us the right to your data” as a condition of booking a flight. The platforms that build trust through transparent, traveller-controlled data architecture will have a structural advantage in adoption – and adoption is still the problem that kills most travel technology innovation, AI included.
Privacy architecture is not a compliance layer. It is a product strategy. The agent that a traveller trusts is the agent they use. The agent they use is the one that gets better data. Better data produces better outcomes. Better outcomes build more trust. This is the flywheel the industry should be building – and it starts with giving travellers genuine control over their own memory.
The infrastructure question is being answered by default
The GDS did not emerge because airlines sat down and collectively decided to build neutral distribution infrastructure. It emerged because the alternative – every airline maintaining bilateral relationships with every travel agent in the world – became operationally impossible. The coordination problem became undeniable, and someone built the switching layer.
We are approaching an equivalent moment in travel data. The AI investment wave is creating a coordination problem at the profile level that is becoming operationally impossible to ignore: every platform building its own memory layer, every agent flying blind on its own proprietary data island, every traveller entering the same information for the 20th time on the 20th platform, every duty-of-care team unable to locate their people in a crisis because the data lives in too many places to reconcile in real time.
The question is not whether shared profile infrastructure gets built. The question is whether it gets built as open, traveller-controlled plumbing – the kind of infrastructure the industry builds on – or whether it gets captured inside the walled gardens of the two or three platforms large enough to force it into being on their own terms.
The AI agents reshaping travel are impressive. They will get more impressive. But the ones that win will not simply be the most intelligent. They will be the ones that remember – accurately, completely and with the traveller’s trust – who they are talking to.
That requires infrastructure the industry does not yet have. Building it is the most important unsolved problem in travel technology. Everything else is interface.