AI in Travel: Where are we at as of September 2024

For a little over a year now, I’ve been reporting weekly on all the developments happening with AI within the travel industry. To date through the Everything AI in Travel weekly newsletter I’ve looked at over 600 different use cases from all corners of our unwieldy industry including airlines, airplanes and airports to hotels to travel agents and travel management companies.

So where are we at now more than a year and half since OpenAI’s ChatGPT burst onto the scene and threatened to upend everything? In short, I’d say, probably not as far advanced into disruption as I first thought we might be. 

Generative AI is predominantly an order of magnitude shift in the ability of machines to communicate. Travel is predominantly a service industry and a communication-based industry. Of course, we have the hardware of planes and hotels, but it is equally as valid to travel in your own car and stay in an Airbnb. And if we calculate the amount of time we actually spend in the plane or hotel as a proportion of the whole holiday, I hope it is just a fraction. As such, the opportunity is and remains huge to disrupt the way the majority of us travel, for the better.

But we aren’t there. And in many places, it also feels we aren’t as close as we should be. Let’s take a look at some of those spots.

Personalization: there is far more talk about personalization than we actually see in practice as a result of generative AI. On the plus side we have companies like Legends who are helping companies build the customer side data to allow for personalization but there are still too few examples of it happening in practice. 

It feels like a problem that Travel is waiting for someone else to fix. Will consumer side data live on devices via Apple Intelligence for example? (What about Android users?) As a consumer, I like this solution. I tell my phone what it is that I like or don’t and companies connect it with that. The status quo is every company in the world trying to find out about me individually. I wouldn’t even mind my phone asking a few qualifying questions from time to time to make sure it is up to date.

I feel most companies never really got to grips with their CRM. I’m sick of clicking cookie pop ups on every website. We are ready for something better. See below for how agents might help.

Unplanned Travel: There is always a lot of buzz around travel planning because it looks quite obvious as a problem that needs solving. Every new generative AI model release has taken time to show their solution to this problem. It is a classic decision tree problem (with infinite branches) that needs to bring different data points and pieces of information together. 

Founders in this space like to talk about how time consuming and painful this all is. What is talked less about is that for most people, travel planning isn’t a single moment in time. It is many moments across quite a long overall time span. I choose my destination. I look at flights. That might change my destination. I eventually might book my flight and might immediately then book my first hotel. Most OTA’s play in this space. 2 nice and chunky bookings here. 

Even the most anal of planners however still has more decisions to make by the time they hit the tarmac in their holiday destination than they’ve already made. Every day there is the:

1. what shall we do before lunch?
2. where shall we have lunch? 
3. what shall we do after lunch?
4. where shall we have dinner? 
5. what will we do tonight?

These decisions are mostly left to be made on the fly.

My old business Urban Adventures carried 450,000 people a year doing 2 – 4-hour day tours in cities. Half were booked in the last 48 hours. We were filling these gaps for people in the moments they were ready to make these decisions. 

Having the trust of the customer as their guide to make all the remaining decisions easy is the biggest prize left unclaimed in travel. AI will be massive here. Almost no-one is focusing on it. Half of it is food. The other half is experiences. No doubt this is why Kayak founder Paul English started with restaurants in his new startup Deets.

AI Agents: AI agents are AI that can perform tasks on our behalf autonomously. How much of that trip planning we hate is filling out the endless forms to book a trip with the airline, then the hotel and so on? Agents will take care of that guff for us. Alas, right now they are also mostly mythical. 

MultiOn is a company working in this space. They are starting with flight bookings. So something is happening. But not much. 

If we are outsourcing our travel management to agents and want them to use personalization to do that in a way we’ve never experienced before, the question remains how they will find the things that are exactly what we need. 

Current trip planners create general trips for general people. But no-one really wants to be general. They want their specific wants and needs to be looked into and after. Many people are travelling in a group with competing needs which need to be balanced. This is the service human travel agents provide. 

I feel there is still going to be a missing piece here and it is basically the mirror of what Legends are doing with the consumer side. Instead, I feel that work needs to be done on the supplier side. The alternative is for the AI agent to keep rewriting their prompts to zero-in on the specific, most relevant things that will most delight us. Unfortunately, this is also the land of hallucinations meaning not everything we are offered, might actually be real. If the really real things could provide some markers and signposts for the agents to read – then we might actually get somewhere pretty amazing. I feel there is a great ‘picks and shovels’ business here somewhere and it is something we are working on with the Videreo business.

In this world our AI agent can run (virtually) ahead of us and simulate all the permutations of the trip to have a bunch of ready answers to the inevitable unplanned questions and deliver them in a just-in-time manner. That would be personalized and magical.

Anyone with an existing relationship with the consumer by the time they are boarding flight number 1 has this chance to own this space. This includes airlines, OTA’s, big chain hotels, DMO’s but also includes travel insurance, money solutions and eSIM providers. It could also include LLM providers themselves and our device makers. 

Of course, a new intermediary is also possible but that is a huge customer acquisition mountain to climb.