Agentic AI in Travel
Legal and Commercial considerations for those looking to implement Agentic AI into travel businesses.
The biggest new concept in the current Artificial Intelligence (AI) boom is ‘Agentic AI’. While ‘Generative AI’ (such as ChatGPT, Gemini, and others) will respond to prompts by generating text or images, Agentic AI goes one step further by taking action on the user’s behalf in response to commands – the AI acts as the user’s agent in the digital world. In a travel context, this would mean a user instructing the AI to book travel services, and the AI booking and paying for them with the user’s credentials.
While new tools have the potential to improve efficiency and customer satisfaction, they do pose legal challenges:
1. Who takes responsibility when the AI Agent makes a mistake on a booking?
2. Will the AI provide transparency and accountability by explaining its actions?
3. Who will dictate the terms upon which the AI Agent is provided and used?
Key to answering these will of course be the AI’s terms of use. Will the provider negotiate over terms, or will they insist on their standard terms on a ‘take it or leave it’ basis? The user is likely to be in a far better position to secure reasonable terms if they are using specialist travel product. Even then, the terms are likely to limit liability for the AI agent’s actions. Will this limitation be set anywhere close to the potential losses that could be created by the AI’s mistakes? This will be something for any travel company to think about carefully. While Agentic AI will have the power to take action unsupervised, the terms are likely to specify that all actions are subject to user confirmation, putting the responsibility for checking all details on the user. Does this detract from the potential benefits of the system itself?
There are plenty of practical questions too:
1. Where is the AI sourcing the services – through reputable suppliers or unreliable 3rd parties?
2. Is the AI system provided by a travel company, or a generalist with little knowledge of the travel market?
3. Is the AI sourcing options from the full travel market, or only from a narrow selection of providers?
These all call into question whether the AI will be reliable, trustworthy, and provide value for money, key qualities in a booking agent.
Would the AI agent even be certain to provide the best options, or get the best deals for the user? Amazon have recently blocked Perplexity’s AI agent from making purchases on its platform, claiming that it was a breach of their terms of service. Steve Endacott (consumer travel veteran) suggests that this is less to do with their legal terms, and more to do with retaining control of the user experience. Whether or not this is true, it raises the possibility that large travel suppliers, like airlines or hotel chains, might do the same. Ryanair famously fought back against OTAs booking its flights, until it grudging began making deals earlier this year. Would it, or others, seek to block AI agents, and thereby reduce consumer choice for users? Any Agentic AI that does not return results from the full market for travel services will risk limiting choice, and potentially giving poor value for money if cheaper options were available elsewhere.
These legal and commercial considerations should be carefully thought out before committing resources to using a new system. Travel businesses should understand where responsibility and liability lie before things go wrong. As always, do the due diligence before rushing after the ‘next big thing’!
Whether you are a business using AI, or an AI supplier, feel free to get in touch to discuss any of the issues arising out of this article at info@techlaw.co.uk.
► Solicitor
► 0113 258 0033