AI Chat → Booking
The impact of AI Engine Optimization (AIEO) on customer acquisition
As counter-intuitive as it may appear, this article starts off with a disclaimer - none of the content published on our newsletter has been or will ever be written by chatGPT. And that’s because AI, in its current state, can summarize the past but it can’t predict or create the future.
That’s the trick with AI - apply it at the right places for best results. Apply it poorly and the “garbage in, garbage out” rule kicks in. Another dimension especially relevant to the travel, experience & hospitality industry is that it will never be able to replace a friendly smile, a thoughtful gesture or a freshly made bed.
At akin, we’re building an AI agent that can free up operator’s time by providing timely, contextual, multi-lingual responses while guiding customers towards making a booking.
Enough about us, let’s dive in.
Addiction
Customer acquisition has been high on a drug for the past 10 years - Google Ads.
Expedia and Booking together coughed up $7.5 billion in 2022. It’s generally thought that Booking.com might be Google’s largest customer.
With a drug this big, the downstream implications are significant. Most companies have the following
People - Highly specialized operators adept at pouring money into the Google engine and playing the adwords game
Software - Specialized tooling and funnels designed to monitor performance and optimize for clicks
Habit - Processes around incentivization, training, content creation, A/B testing and numerous other functions all in service of Google ads driven monetization
Together, they lock incumbents into a mindset of straightening deck chairs on the Titanic. From Kodak to Nokia to Xerox, history is littered with giants stumbling during periods of technological transformation.
AI rocks the boat
Over the past 6 months, the world has come to realize that AI has come of age. Google is scrambling to adapt and OpenAI (chatGPT) has taken the lead in what looks likely to be a step function change in how the online world works. Function calls, Llama2, 32k context, AutoGPT and numerous other innovations being released every second week making it quite hard to separate the signal from the noise.
Which of these innovations are relevant to our industry?
Let’s start by breaking down travel requests into their core components and evaluating the impact of AI on each piece
1. Search and discoverability
LLMs (large language models - commonly referred to as AI) have digested large volumes of data available on the web and are better suited to answering the question above than Google. Marketplaces are currently scrambling to incorporate AI-search into their offering. Meta-search would always be more valuable to the customer and because it’s not feasible for one marketplace to have all encompassing information. A meta-tool like chatGPT or Google (AI) would be infinitely more knowledgeable. Brand-related loyalty and habitual behavior (Airbnb + Booking) might offer some defensibility to marketplaces for a while, but customers are ruthlessly shopping around for better prices and this will not last forever.
Winner - Customer
2. Rates and availability (API)
This information changes every second (increasingly generated by pricing AI) and cannot be memorized by LLMs (chatGPT). So an API to real world information is necessary. Assuming the AI exists to help the customer, it follows that AI would choose the cheapest price to book. And if that is true AND operators are able to connect their price engines to AI engines, it could very well be game over for the middleman — Booking.com, Expedia, etc. But don’t expect them to go down without a fight. The game will quickly evolve into rewards, promotions, discounts, hidden fees, non-refundable rates, etc.
Winner - Customer
3. Product Information
Price alone isn’t a deciding factor. Attributes such as proximity to the beach or wifi speed could be important to the customer. There are two sources of product information
Travel Operator - Real world information that changes all the time. Much of it is offline but relevant (Room 101 overlooks the pool). Some information takes time to show up in the online sphere (if at all).
Reviews / blogs - multiple 3rd party sources would be more believable on subjective questions such as cleanliness, friendliness, etc.
Will AI be legally allowed to live-scrape marketplace reviews? What if the product doesn’t match online promises? How would customers get insured against misrepresentations? How will chargebacks be handled if the AI hallucinated?
AI will be able to crunch multi-variate information quickly and will be able to match customers with their desired preferences. Traditionally marketplaces have charged a premium for this matching function and as AI becomes ubiquitous, marketplace matching wouldn’t be a sufficient differentiator. However, there are lots of moving parts here.
Winner - Operator
4. Historical booking preferences
Customers have certain implicit preferences that might not be apparent to them or that they might not explicitly call out. In this case, booking history + customer reviews (across multiple trips) becomes a huge decider in finding the right option on a subsequent trip. Currently marketplaces have a good overview of customer behavior and are well poised to provide the best recommendations assuming customer behavior hasn’t changed (having a kid drastically alters travel preferences). Switching costs would be quite high.
Winner - Marketplace (Airbnb, Expedia, Booking.com)
5. Offline customer preferences
Online booking behavior is only the tip of the iceberg. The real magic happens when operators deliver the booked experience. Guests might cancel, extend, request for add-ons and reveal a multitude of preferences that were previously unknown. Can hospitality operators capture this data and then use it to lure back customers when they’re ready to book again? Loyalty programs, upgrades, promotions, discounts would fall here.
Winner - Operator
The Bottom Line
Customers
It should come as no surprise that the winner in all of this is the customer. AI will be a commodity and all the surplus of AI related economic gains will go to the customer. What that means in the travel context is that customers will be able to book better holidays for cheaper. Better as in better suited to the guest’s needs and expectations. And AI will not play tricks to get people to pay a higher price for the same offering.
Marketplaces
The creation of Google Adwords coincided with the emergence of Booking.com as a superpower who were relentless with A/B testing and funnel optimization. A step change in the technology underpinning search could result in a few winners and losers. Any bets?
Search
Is Google going to remain undisputed? Many observers have likened this technological revolution to the browser wars of the early 90s. Microsoft won against Netscape because they could package Internet Explorer with Windows. Will Google be able to pull off the same trick? User habits (“let me Google that”) are notoriously difficult to break. On the other hand, Google is struggling to get started on the AI game. Will this fumble result in their demise or is the upcoming Gemini AI release going to save them?
Operators
Whether it’s transportation, accommodations or experiences, travel operators have an opportunity to fight back for their fair share of direct bookings. And the key decider will be data. Can they get the act together and collect meaningful offline data, coupled with targeted promotions to lure AI agents into making direct bookings?
Bring out the popcorn folks. This is going to be turmultuous and wildly entertaining.
Nobody, not even AI, can see where the dice are going to land. I, for one, am very excited about where this road leads us because maybe the golden age of experiences, hospitality and travel lies ahead of us and not behind.