The Long Way Around to CollabsMap

Building a startup is an exercise in first finding out what people don't want, on the journey to discovering what they do. Here are a few things I've learned people don't want.

DMOs don't want this:

A searchable map of their destination that automatically and continually populates with videos from creators they've already engaged, as soon as that creator posts to Instagram or TikTok.

Each video showcases something unique you can actually experience in all corners of the destination, with direct links to book with the local hotels, tour operators and attractions providing those experiences. OTA’s don’t even have this merchandising.

We thought we’d solved a bunch of big problems here! I thought I’d be competing with Trump for the Nobel Peace Prize having finally brought DMO’s and their suppliers together in joyous harmony!

• Direct Bookings for suppliers rolling in.

• Supplier ←→DMO relationships solved.

• ROI on DMO creator marketing measured in local bookings.

Definite hard no on this one. 🙃 (Suppliers we up for it tho..)

Content creators definitely don't want this:

We built the tech to take a creator’s entire back catalogue of travel videos and automatically turn them into hundreds of bespoke product pages featuring the exact hotels, tours and experiences they’d personally tried. 400 videos = 400 products. No extra travel required. Live in 2 minutes.

All that was required was to re-post the video and point people to the product link.

Don't be so stupid, Tony. Why would anyone want that? One (1) creator out of 200 did anything.

And travel brands definitely didn’t want this:

A dashboard showing exactly what happened after they spent money on creator marketing.

Views. Engagement. Leads. Clicks. Bookings. Revenue.

Actual ROI, with full transparency.

God forbid. (I mean, what if the dashboard showed that they were wasting their time and money?)

I know when I held the P/L reins, I definitely wanted this. I was bullish here.

The funny thing is, all three still seem like perfectly logical bets to me. Even sitting here now, knowing for sure that nobody wanted them badly enough.

We kept trying to close the creator-to-booking gap

That's really what the journey of Videreo has been about. Creators are extraordinarily good at creating travel intent.

Someone watches a Reel of Palau and thinks: I need to go there. Someone sees a creator eating their way through Rome and saves the video. Someone watches a hotel walkthrough and sends it to their partner.

Demand has been created. But then something strange happens. The creator disappears from the transaction.

Rafat Ali described this problem brilliantly this week in Skift's The Broken Creator-to-Booking Chain, and Why It Matters Now.

His example is simple: a tourism board spends $200,000 on creator videos about Mexico City. Two million people watch. Someone saves a hotel they discover in the content. Three weeks later, they're sitting at their laptop, Google the hotel and book it through Expedia.

The tourism board paid to create the demand but can't prove it produced the booking.

The creator caused the inspiration but doesn't participate in the transaction.

And Expedia captures the sale despite having nothing to do with creating the original demand.

That is the broken creator-to-booking chain.

And we spent a couple of years trying to fix it.

We were trying to solve a problem several steps further down the funnel than many customers were ready to solve.

Then we found something people actually wanted

While we were busy trying to reinvent creator commerce, we had accumulated something much simpler.

Creators. Thousands of them.

And on the other side were travel businesses saying: "We'd like creators. How do we find the right ones?"

So we built CollabsMap. Put an opportunity on a map. Let creators discover it. Let them apply. Let the travel business choose who it wants to work with.

No grand reinvention of travel commerce required. And people used it.

Today CollabsMap has thousands of travel creators discovering opportunities from nearly 1000 travel brands around the world. The basic marketplace is free for creators and deliberately easy for travel businesses to join (freemium).

And that simple utility has opened the door to something much bigger.

Perhaps we'd been solving the problem backwards

The Skift article identifies five emerging creator business models in travel: flat-fee advertising, affiliate commerce, in-platform commerce, creators acting as travel advisors or sellers, and large audience-access media deals.

They all represent different attempts to answer essentially the same question:

How does the person who creates the demand participate in the economic value of that demand?

I still think that question matters enormously. But perhaps the first job isn't to solve the entire creator-to-booking chain. Perhaps it's to build the infrastructure where travel brands and travel creators can reliably find each other and work together in the first place.

That's CollabsMap.

And once you have that network, you can start adding the other pieces back in.

Better campaign management.

Guaranteed distribution.

Performance measurement.

Lead generation.

Affiliate links for exact ROAS

Booking attribution where it's possible.

Paid amplification of the content that organically proves it can perform.

Suddenly all those things nobody wanted aren't necessarily bad ideas. They may simply have been in the wrong order.

The timing is getting interesting

There's another reason this matters now.

AI is making generic travel content extraordinarily cheap to create.

Skift argues that this could make genuine human trust and recommendation more valuable at exactly the same time that monetising that influence becomes more complicated.

The internet isn't going to suffer from a shortage of perfectly written articles about 10 Things To Do In Paris. AI can generate those forever.

What remains scarce is a real human being standing somewhere saying: "I've actually been here. I did this. And it was bloody good."

Travel creators already generate that signal at enormous scale. The industry just hasn't built particularly good infrastructure around it yet.

We've spent the last few years discovering several ways not to build that infrastructure. Which, as it turns out, might have been exactly what we needed to do to finally understand how to build it.

And that's how Videreo became CollabsMap.

The creator-to-booking framework referenced throughout this article comes from Rafat Ali's excellent September 15, 2026 Skift article, “The Broken Creator-to-Booking Chain, and Why It Matters Now.”

Read the original Skift article

This content is provided by the newsletter sponsor CollabsMap.com - find the creators who are ALREADY heading your way. Put an offer to them. Choose the ones you want to work with in a click.

The AI trust problem

 AI may start the trip but it is still a long way from closing it.

New research from Mower found 28% of AI-using US leisure travellers would start planning their next trip with an AI tool. Google search was only just ahead at 30%.

But here is the interesting bit. After AI recommends something, 59% go back to Google, 49% visit the official website, 40% compare prices across multiple sites and 20% ask another AI tool. Just 5% book exactly what the AI suggested.

So AI discovery is not replacing the rest of the funnel. At least not yet. It is adding another front door, then sending travellers off to fact-check what they found.

For travel businesses, that means two jobs now. You need to get into the AI answer, but you also need to survive the verification lap. Current prices. Clear policies. Useful reviews. Accurate listings. A direct website that does not look like it was last touched in 2014.

AI visibility without trust will get you mentioned, not booked.

A couple of caveats: this is Mower's own research, distributed as a press release, and the 1,000-person sample deliberately included US leisure travellers who already use AI at least occasionally. So it does not tell us how common AI travel planning is across all travellers.

But I think the behaviour is the signal. Recommendation and transaction are still different problems.

If an AI recommended your business today, would the traveller find enough proof to keep going?

Get Your Guide’s solution to the trust problem.

GetYourGuide has been testing a Creator Review program for tour operators with new or under-reviewed listings. The idea is pretty simple: connect operators with creators who take a free booking, then leave a detailed, photo-backed review within 24 hours. Skift says GetYourGuide's own research found travellers are three times more likely to book an activity once it has three reviews.

If you run tours or experiences, you already know the pain here. A new product can be good, the photos can be decent, the guide can be excellent, but if the listing has no reviews it is asking the customer to take the risk. And customers, quite reasonably, do not love taking the risk when there are ten other options with proof sitting right next to yours.

GetYourGuide is also rolling out AI assistant integrations with Gemini, ChatGPT and Claude, plus AI-assisted operator analytics and content suggestions. That is interesting, but I think the more important story is what those AI layers will need to trust. Clean product data helps. Good content helps. But marketplace proof, recent reviews, photos, labels and structured signals are going to matter a lot if AI is deciding what to surface, summarise or recommend.

The slightly awkward part is that this is still paid-for with a free booking, even if the review is labelled and not guaranteed to be positive. So it needs transparency, otherwise it just becomes another version of influencer mush.

But the direction of travel is worth watching. Creator content is not just fluffy awareness if it helps a listing get found, trusted and booked. It becomes part of the conversion plumbing.

If you are an operator, I would be asking a very boring question: what proof does your product have today, and is it easy for a human, a marketplace algorithm and an AI assistant to understand it?

What do you think? Useful or does the free-booking review model make you twitchy? Can you solve a trust issue with something that might not pass the pub test?

Is AI making the passage into the airport quicker and smoother?

Salesforce has a new case study on TSA's AI agent, Ace, and yes, it is Salesforce news so apply the appropriate vendor-story salt. But the numbers are still worth a look.

Ace is apparently handling around 100,000 traveller conversations a month and resolving 96% of routine questions without escalation. Things like liquids, medical devices, checkpoint prep, busy travel periods, and all the small questions that do not feel small when you are stressed, late, and standing in an airport queue.

The useful bit here is that TSA did not start with the shiny public-facing agent in isolation. According to the piece, it first built the plumbing: a customer service manager app, a TSA Cares app, a central place for traveller interactions, and then the AI layer on top of that. The agent works because there is some operational structure underneath it.

That is where I think a lot of travel AI will either work or fall over. Hotels, airlines, tour operators, airports, DMOs and OTAs all have endless repetitive customer questions. AI can absolutely take a chunk of that workload. But only if the answers are accurate, the escalation path is clear, and the data sitting behind the agent is not a mess.

Salesforce says the wider TSA applications are projected to save more than 11,660 staff hours annually and $11 million in labour-based value over three years, with consumer interaction costs down more than 90%.

That is not demo theatre. Or at least, it is a much more useful kind of demo theatre.

For mine, this is the operator lesson: AI customer service in travel is not really about replacing people. It is about protecting human time for the moments where judgement, empathy, risk, refunds, disruption, accessibility, security, or weird edge cases actually matter.

And travel has a lot of weird edge cases.

If you are building AI into your service flow, the question is probably not "can we launch an agent?" It is: what boring, repetitive, high-volume work can we safely remove from the team, and what data/knowledge plumbing has to be fixed before the agent can be trusted?

AI as a partial solve to customer service should no longer be news. Airbnb banged this drum for 18 months. But if its news to you then, well…………

BREAKING: Booking talks about the connected trip

Booking's AI story is not really a chatbot story. It is an acquisition-cost story.

Yahoo Finance has a short piece on Glenn Fogel talking up Booking's AI-powered Connected Trip strategy. Usual investor-conference caveat here: companies are always very excited about the thing they are already building. Funny that.

But the useful part is the commercial logic underneath it. Booking does not just want AI to answer travel questions. It wants AI to make it easier for a traveller to book the hotel, flight, airport transfer, attraction and restaurant inside the same ecosystem.

The dream is owning more of the trip without having to reacquire the same customer five times.

That is where this gets interesting for travel operators. If AI makes connected trips easier, the winners are not necessarily the businesses with the nicest chatbot. They are the businesses with the supply, payment, loyalty, support and data plumbing to turn one booking into the next booking.

Booking has been pushing this for years, so the idea itself is not new. What is changing is that AI gives the connected trip a more natural interface. Instead of forcing travellers through separate tabs and filters, the agent can hold the whole mess together: dates, budgets, preferences, weather, tickets, timings, transfers, cancellations and all the annoying little decisions that make travel planning feel like admin with nicer photos.

Everyone knows Booking is trying to achieve this. They’ve been banging on about it since Moses was a boy. Human travel advisors achieved the connected trip booking in the 70’s. No-one in the internet age made it.

In AI age it seems there is a new chance but there is one piece I think that gets nearly every AI builder out there stuck. AI makes everything faster. It compresses everything into an instant “obvious” solution. But humans are slow to spend $4K. They want to be sure, really sure, they are making the right decision and getting the right price, on just about every individual piece. Speed of answer currently does not give most that confidence.

My read: connected trip is where a lot of the serious AI in travel action is going to sit. Not because it can pull all the threads together at lightning speed and jam a “book now” button in front of people but rather a system that gives the human choice autonomy and then does the boring bits of price shopping and review comparison to land on the final decisions. That’s decisions with an s on the end. The connected trip will not be a single decision on a full package. It never has been and people don’t want it to be. It is a combination of many micros decisions made over time.

And it doesn’t need to be just an OTA solution. If you run a hotel, tour, DMO or travel brand, I would be asking a fairly boring question this week: could an AI agent actually work with your business cleanly, or would it get stuck in the booking plumbing? Agents are coming to ask what you have to offer. You need to ready with a bunch of different answers to get the booking.

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:

New ad format inside GPT

ChatGPT ads are starting to look less like ads with a link to click and more like sales agents to nurture you through to purchase.

Juozas Kaziukėnas spotted a Wayfair ad inside ChatGPT where the click did not send him to Wayfair’s website. It opened a branded Wayfair GPT with a pre-filled shopping prompt and ability to continue to ask all the questions you need answered until you have the confidence to hit BUY

If this becomes a proper format, the ad is no longer just the creative. It is the agent. The product data, availability, policies, FAQs, booking rules, cancellation terms, room types, tour inclusions, upsells and payment flow all become part of the media buy.

Which sounds exciting. Also slightly terrifying. Because if you are a hotel, tour operator, OTA or DMO, the question is not just “can I advertise in ChatGPT?” It is “what happens if the customer never comes to my site at all?”

The ad click becomes a conversation. The conversation becomes the funnel. And whoever controls that conversation is sitting in a very useful position.

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.

This week I chatted with Thomas Bangert, founder of Travecta, about building the connectivity layer he believes multi-day travel has been missing.

This could be your brand’s newsletter (and podcast).

Partner with Us - sponsorship info for the newsletter and podcast can be found here.

Most clicked last week was the link to the piece on Claude’s new commerce agents.

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