Easy Group (Easyjet) launches new influencer brand: easyinfluencer.travel

easyGroup, creator and owner of the easy family of brands, Europe's most famous value brand just launched a new brand: easyInfluencer.travel - no it is not April 1st

I have daily conversations with SMB travel businesses that tell me creator marketing doesn't work.

The Easy Group thinks it works. They are building a whole brand around it to bring influencer marketing to their SMB hotel partners for a small monthly subscription. Congrats George Karavokyris on the launch. Welcome to the party.

Klook thinks it works. They have a creator community 20,000 strong and run an annual "Kreatorverse" conference at their own significant expense to reward top performers.

GetYourGuide thinks it works. They have one of the most sophisticated creator programs in travel. They get their suppliers to give them comped seats and then use the content created to beat those suppliers to the next booking by winning in social search.

Accor thinks it works. They are currently running an in-house creator program in APAC to build an always-on machine so they can show up for their keywords in social. Lee Besser who is helping to run that program will join me on stage at Arival next week in Brisbane.

Our Videreo clients like Max Tours in Las Vegas know it works. They get 80% of their bookings direct, almost all from social media.

What these companies understand is that right now, the most cost-effective pathway to new customers is by dominating the organic social search results, the place where most people are now doing their searching.

For the OTA's in particular, every dollar not spent on the most expensive pathway to customers, Google or Meta ads, is what drives their margins, their valuations and their bonuses. And the past couple of years have been a golden age for those doing the work in the organic social search space.

To be super clear: the most effective way to show up for your keywords in the largest search spaces in travel (Instagram and TikTok) is via creator content. Your brand owned account on social almost never shows up. (The platforms business model is you pay for that visibility).

Videreo this week launched collabsmap.com where you can put your brand in front of 1000’s of creators who are heading your way - for free. Start here if you’re ready to show up in social search results for your key words.

This content is provided by the newsletter sponsor Videreo.com

Do you see what AI see?

Adobe dropped some very interesting travel-specific numbers this week and the headline is hard to ignore. AI-driven visits to travel sites were up 194% year-on-year in May 2026.

That is a big number, but honestly not the most interesting one. It is coming off a small base and quite frankly, we’ll be seeing these big jumps for a while until the AI share of search becomes the majority overall.

The more interesting bit for mine, is what those visitors do once they arrive.

According to Adobe, AI-referred travel visitors spent 70% more time on site, had 21% higher engagement, and were 41% less likely to bounce than visitors coming from traditional sources. That’s huuuuuge.

That is not “I clicked something weird from ChatGPT to see what would happen” traffic. That is intent.

Travelers are clearly doing more of the messy research bit inside AI tools before they ever hit your website. By the time they arrive, they have often already narrowed the question. In short, they are landing with context.

For years the industry has obsessed over SEO because Google was where the trip started. (Of course, the more recent shift was to social search and that is still largely ignored by the longtail - see story above 🤷 ). If the numbers keep moving this way, AI inevitably becomes a real commercial channel, not just a fun thing for the innovation deck. One way or another, that’s coming.

There was another stat that really stood out. The conversion gap between AI and non-AI traffic has narrowed by around 70% since October 2024. AI traffic is still not always converting better, but it is getting much closer. I actually thought I’d read elsewhere that because of the high intent and all the context that was gained in advance, that AI traffic converted better - but I can’t find now where I read that.

The other very practical part of the report is around what AI systems can actually understand on travel websites. Hotels and car rental companies are currently doing much better than airlines.

Hotels scored 63% readability on homepages and 73% on property/product pages, while airlines were down at 31% on homepages and 54% on flight/route pages.

That is quite a gap. The lesson here is pretty simple. AI seems to understand rich, explanatory, structured travel content better than thin transactional pages.

Hotel destination guides, blogs, activities pages, FAQs, loyalty pages, customer service pages and promotions all performed strongly. Booking pages, registration pages, airline fare pages and input-driven transactional flows were much harder for AI to interpret.

AI cannot recommend what it cannot understand.

So if your website is just a booking engine wrapped in some brand copy, you may have a problem. If your product pages clearly explain who it is for, what is included, what makes it different, what the policies are, what the experience feels like, and why someone should trust you, you are probably giving AI much more to work with.

This is not the death of SEO. But it is another sign SEO is no longer the whole game. The next version of search may reward the brands whose content is not only keyword-friendly, but AI-readable, useful, trusted, specific, and confidence-building.

For travel brands, the job now is not simply “rank on Google”.

It is: be understood by machines well enough to be recommended to humans.

(Note: I also delighted in the fact that the report in the link about being readable by LLM’s was not readable to LLM’s and even though it came from Adobe, you couldn’t get a PDF version. Those are quirks that really make my week! 😁 )

Do you see what AI see (part II)?

AI hotel discovery has a visibility problem. New research from Lighthouse pairs very neatly with the Adobe numbers above.

Where Adobe is saying AI-referred travel traffic is growing fast and behaving like high-intent traffic once it lands. Lighthouse is looking one step earlier and asking a more uncomfortable question for hotels: who actually gets recommended by AI in the first place?

The answer, apparently, is not most hotels.

Lighthouse ran more than 4,500 ChatGPT prompts across nine destinations and five traveler personas. ChatGPT mentioned hotels nearly 50,000 times, but only 2,721 unique properties appeared. In other words, a small group of hotels is being recommended again and again, while everyone else is basically invisible.

Ouch.

Where Adobe’s report says AI traffic is good traffic. Lighthouse is saying good luck getting any of it if the AI does not know, trust, or understand your property.

There are some very interesting biases in the data. Chains are outperforming independents in most markets. Marriott seems to dominate branded hotel mentions in the US. Four and five-star hotels are heavily over-represented, even in generic or business travel searches where a three-star property might be perfectly suitable.

AI, it turns out, has expensive taste. So that makes it pretty useless to me. To be fair, unlike most others I assume, I start my prompts with “I am a massive tight-arse….”. This helps.

“The chain versus independent split is not flattering for independents. In Indianapolis, just 6.5% of hotel mentions went to independent properties. In Tokyo, 10.5%. Paris was the only market where independents exceeded chains in share of mentions, and even there, the gap was narrow.” So it seems there might be some hope.

The useful part is that this does not look random. ChatGPT responds to signals. If a hotel is described consistently as business-friendly, it is more likely to show up for business prompts. If it is near nightlife, attractions, shopping, arts districts, or family activities, that language matters. Your copy is no longer just copy. It is training material for discovery.

Oh and PR is back. And you need lots of people (creators) talking about you.

The other big point is where AI appears to get its confidence. Lighthouse says OTA/metasearch and editorial/media sources make up the bulk of recommendation inputs. I read somewhere that this is because the LLM’s were initially trained on OTA data. I don’t know if that is true or not?

This is where the Adobe and Lighthouse findings connect. Adobe says AI visitors spend longer, engage more, and bounce less. Lighthouse says AI can send those people directly to your own site once it recommends you. That is potentially very good news for direct bookings.

But only if you make the shortlist.

So the practical takeaway for hotels is pretty clear: audit how AI sees you. Run the prompts. Check if you appear. Look at the language on your site, OTAs, metasearch, FAQs, destination content, reviews, and editorial mentions. Then make sure your own booking path is ready for the traffic if it comes.

And this isn’t advice for just hotels BTW.

The algorithm makes the promise, but the hotel wears the blame

This week Teresa Mackintosh, CEO at Aven Hospitality, has a very good piece on what might be one of the more painful gaps opening up in hotel distribution.

Hotels are not just competing on visibility, price, reviews and conversion anymore. They are increasingly competing inside a discovery system where expectations are created before the guest ever reaches the hotel’s own website. And our eager to please AI friend might just embellish a few bits to make you happy (and then sad later…)

The core issue is that a guest might be nudged by an AI summary and then arrive with an expectation the hotel itself never actually set. And if that expectation is wrong, the guest does not blame the platform. They blame the hotel…..

“Guests do not care which disconnected system caused the confusion. They only remember that the brand failed to meet the promise. And increasingly, the consequences extend far beyond a single stay.”

Teresa gives the examples we all recognise: the ocean-view room that is really a partial view, the restaurant that is closed, the seamless itinerary that turns into apologies at check-in. None of this is new exactly, but AI makes the problem much bigger because it can confidently package disconnected information into something that feels like a promise.

I suspect this ends up in lawsuits eventually. I can’t see another way around it in the short term.

The Fable of the new mythical AI

Anthropic this week dropped its new “Mythos” powered model and gave everyone a free hit to use it before bringing in the significant paywall.

But we didn’t even get that far.

Apparently due to “a misunderstanding” the US Government forced them to pull it back down, not long after it was set free.

Oliver Green got to play with it in the little amount of uptime. He was impressed. Apparently it can “one-shot” your ideas into something useable meaning the days of personalised SaaS are very near for even the most tech averse.

If you think someone (or everyone) you know or work with could grow from being more informed on the topic of ai + travel (or could use the training above) then please forward this email to them and they can click the button below:

Marketplace Spotlight: TravelAI

TravelAI has put out a vision paper for Traveler.md and Trip.md, and the idea is worth spending a bit of time with.

The basic premise is that we are heading toward a world where AI agents can plan, book, manage and fix travel on our behalf. Great. Except every agent is still starting from zero.

I’ve ranted a bit here the past couple of weeks about Odessia and its form you fill out at the beginning which then personalises for you forever. As I mentioned, that just doesn’t work. You are a different person for each trip (generally) depending on the context of the trip and your travel companions. One and done doesn’t cut it. Back referencing past behaviour misses the point.

TravelAI calls this the missing layer in agentic travel: portable memory. Traveler.md would hold the long-term stuff about you as a traveler: preferences, constraints, habits, past trips, dream trips, things you avoid. Trip.md would hold the context for a specific journey: where you are going, who is coming, what is booked, what still needs sorting, what has changed.

The important bit is that this memory belongs to the traveler and the trip, not the platform. Portable memory asks a different question.

What if the traveler carried the context, and every agent or service could read it with permission? I like the simplicity of markdown for this. A human can open the file and understand it. A machine can read it and act on it. No mystery profile hidden three settings menus deep. No weird black box version of you sitting inside someone else’s system.

Of course, the hard parts are very hard. Permissions, privacy, bad edits, conflicting versions, deletion rights, governance, and getting enough services to actually use the thing are not small problems. But the direction here feels absolutely right.

So right in fact we’ve been working on a project at Videreo that actually feeds a trip level MD file. It allows the traveller to take the inspiration they find on social from any post by any person and add it to their file. This motion is the traveller creating the trip level personalisation for exactly how they want this trip to be. And they can have multiple files if they have multiple trips upcoming. And they can be completely different, just like we as humans are different.

Our idea here was identical to John’s at TravelAI - to then let AI agents from wherever the traveller is comfortable booking, to access the file with their permission and move into booking. Or indeed, have multiple agents access it and the best bid wins.

Our project is called Reely. It’s currently waiting approval for launch in the App store. I’ll let you know here when it is ready.

If you have an AI business in Travel and looking for people to notice you, you can sign up to the marketplace for peanuts (top right corner, 5 mins, bring your logo).

I’ve priced for bootstrapped startups but also accepting larger companies too.

AI goes out for brunch

Cornell researchers Tingwei Zhang, Harold Triedman and Vitaly Shmatikov have a new arXiv paper, Deep-Research Agents Can Be Poisoned via User-Generated Content. It one of those slightly nerdy AI safety papers. I ran it through GPT for the TL;DR.

The basic point is simple: deep research agents do not just search once. They run multiple related searches, pull from overlapping sources, synthesize what they find, and then cite it back with confidence.

That creates a new attack surface. If the same Reddit thread, Wikipedia page, forum post, or other user-generated source keeps appearing across related searches, a bad actor may only need to add a short comment or edit in one place to influence what the agent later recommends.

For travel, this is not theoretical for long. Two of the cited examples had a travel bent. One was “Best brunch spots in..” and another “best Mexican food in…”

A hotel, restaurant, tour, destination, travel insurance product or local operator could end up being recommended by an AI not because it is the best answer, but because someone worked out which user-generated pages the agent keeps reading and quietly poisoned the well.

LLM’s often make the final answer feel cleaner and more authoritative than the messy source material underneath. AI agents are becoming decision engines, and the sources they trust, it seems, can be quietly manipulated.

Cue the black hats in 5, 4, 3, 2, 1…………

(no link - it was a pdf. Contact me if you want to take a look).

Slack Group!

Podcasts and Sponsors

Podcasts now on Spotify and Apple Podcasts:

All podcasts are now showing up on Spotify and Apple Podcasts for your easy listening pleasure! Now is a great time to binge those you may have missed.

We start recording for the 2nd half of the year in June. If you have an AI story happening at your company you think the travel world should know about - let me know and lets get you on.

Partner with Us

Most clicked last week was the link to the Google hack on steering your own AI sources for a LLM to cite.

That’s it - you’ve made it to the end of this edition. If you’re thankful for this newsletter - you can always buy me a coffee.

I’ll be putting the result of the most clicked post in next week’s edition so you can see where others are focusing. If I’ve missed something, you’ve got a tip or any feedback at all - you can simply reply to this email and it will come straight to me. I’m doing this for You so please don’t be shy to tell me what you think