Holly Lewis

13 Nov 2026

BrightonSEO came to the San Diego Convention Center on September 15 and 16, 2026. We went through 62 speaker decks from the event to pull out what actually matters for hotels.

Very little of it was hospitality-specific. Most of the data came from retail, SaaS and ecommerce. But three themes kept coming up, and each one lands squarely on hotel marketing:

  • Fewer searches end in a click, and AI now sits in the middle of the booking journey.
  • Third-party mentions shape AI answers far more than your own website does.
  • Many AI crawlers can’t read content that loads with JavaScript, which is how many hotel sites show rates and rooms.

One caveat before we start. Most figures shared at the event came from vendors or single case studies. We’ve treated them as signals worth testing on real hotels, not settled facts.

Fewer Clicks, but AI Isn’t Replacing Search

The click is shrinking. SparkToro research based on Similarweb data found that 68% of US Google searches ended without a click in early 2026, a figure Jen Cornwell of Tinuiti highlighted in her talk. Ahrefs’ updated study, presented by Rob Delory, found that an AI Overview on the page now correlates with a 58% lower clickthrough rate for the number one result. When Ahrefs first measured it in April 2025, the figure was 34.5%.

But AI isn’t replacing search. It’s sitting in the middle of the journey. Baruch Toledano of Similarweb showed that US shopping journeys involving AI grew from 4.5% to 11.4% over two years, and 89% of those journeys also used search. Journeys using both converted at 23.0%, against 17.3% for search alone and 12.5% for AI alone, a pattern also reported in Similarweb’s State of Ecommerce 2026 report.

That’s retail panel data, but Toledano’s travel example showed the same pull. When AI recommended Kayak, 12.0% of users went on to visit Kayak, against 3.4% who visited Skyscanner. Being the name AI gives sends people to you.

Hospitality discovery is even more spread out. Damian Rollison of SOCi showed that only 34% of hospitality searchers start on a search engine, the lowest of the eight categories SOCi tracked. 17% start on AI, 14% on social media, 13% on industry-specific sites, 12% on review sites and 11% on maps. SOCi’s 2026 Local Discovery Index, a survey of more than 1,000 US consumers, found that AI use for finding local businesses jumped from 9% in 2025 to 52% in 2026.

What it means for hotels: AI visibility often shows up later in the journey, as branded searches and direct visits, not as AI referral traffic. Clicks alone will understate how visible your hotel really is. Judge SEO on impressions, branded search and direct bookings alongside traffic, and plan social, search and web together rather than as separate channels.

Third Parties Write Most AI Answers

This was the clearest message of the event. Nikki Lam of NP Digital showed that 82% of citations on commercial prompts came from third-party sources. Just 3% came from the brand’s own site.

It matches Ahrefs’ study of 75,000 brands, cited by Michael Johnson. Branded web mentions correlated with AI Overview visibility at 0.664. Backlinks managed 0.218.

Hotels do have a little more home advantage than most. Chris Donnelly’s citation data put the median share of a hospitality brand’s AI citations coming from its own site at 33.7%, against 19.5% in health. Your website still counts, but around two-thirds of the fight happens elsewhere.

For a hotel, “third parties” means OTAs, review sites, travel press, “best hotels in” listicles, Reddit threads and YouTube. Two of those stood out:

  • Reddit. Ross Hudgens of Siege Media found travel gets more value from Reddit than any of the other nine industries he compared, at 2.31% of site value against a 0.99% average.
  • YouTube. Jen Cornwell, citing Otterly.AI, showed that long-form videos make up 94% of YouTube citations in AI search. View counts didn’t predict which videos got cited.

Sebastian Pawlowski of Lastmile Retail showed how this plays out for hotels. His team analyzed 137,455 local search results pages in New York, Chicago and Los Angeles. A search like “marriott hotels near broadway nashville” sends people back a step, with Google asking “Is it better to stay near Broadway or Music Row?” Marriott appeared on 47.8% of its purchase searches but only 16.3% of the follow-up questions they generated. Hilton dropped from 25.5% to 7.9%. Reddit went the other way in hospitality, from 22.9% to 70.3%.

What it means for hotels: another page on your own site won’t move AI visibility much. Earned mentions will. Find out which sources AI already cites for your hotel and your competitors, then aim PR at those. Turn unlinked brand mentions into links. Pitch data-led stories to local and trade press. Brandon Leibowitz made the case for guest podcasting as an SEO channel. For hotels, that means GMs and chefs on local travel and food podcasts. And keep an eye on your city’s subreddits, which is where social listening pays off.

This is the point where SEO and PR stop being separate jobs.

Can AI Actually See Your Rates?

Before any content work, check that AI can read your site at all. Serge Bezborodov of JetOctopus live-tested AI user bots on stage and concluded they are “simple text fetchers.” Chris Donnelly made the same point: non-Google crawlers still struggle with client-side JavaScript, so a site Google can render may stay invisible elsewhere.

Speed matters too. Zach Chahalis shared a Profound analysis of 700,000 pages. Pages that often timed out when AI crawlers tried to fetch them got 18 times fewer citations, and pages that failed more than three-quarters of the time got none at all.

That’s a real problem for hotels. Rates, room details, offers and amenities often load inside booking widgets and scripts. If that information isn’t in the page’s raw HTML, many AI tools simply can’t see it. They fill the gap from OTAs and review sites instead, and they don’t always get it right. In SOCi’s survey, 67% of AI users said they’d been given wrong information about a local business at least once, and 63% of consumers had chosen another business because of missing information.

The details matter at the moment of choice. For hospitality, Rollison found offerings and deals (36%) and location and hours (35%) ranked just behind reviews (40%) as the factors that decide where people book.

A basic AI access check takes less time than you’d think:

  • Compare the page source with the rendered page on your rooms, offers and amenities pages. Anything that only appears after rendering may be invisible to AI.
  • Check robots.txt isn’t blocking the AI crawlers you want to reach your site.
  • Check your CDN and firewall bot settings. Security rules can block AI crawlers without anyone noticing.
  • Look for AI bot visits in your server logs.
  • Ask ChatGPT, Gemini and Google’s AI Mode about your hours, parking, pet policy and amenities. Log what they get wrong.

The fix is usually simple in principle: keep your key facts in plain, server-rendered text, not only inside a widget.

Track the Consideration Set, Not the Rank

Measuring AI visibility is still shaky. Nicolas Rabouille cited research from SparkToro and Gumshoe finding less than a 1 in 100 chance of getting the same brand list from AI twice, and Rob Delory noted that AI Overviews change every 2.15 days. Thomas Peham of OtterlyAI showed GA4 crediting Claude with 0.1% of conversions, against 10.6% when customers were asked how they found the company.

Since the event, hotel-specific evidence has landed. A study published on Hotel Speak on October 6 analyzed 28,600 answers across ChatGPT, Gemini, Copilot, AI Mode and AI Overviews:

  • The same query, asked an hour apart, returned the same first-choice hotel only 60.2% of the time.
  • Across the five engines, the top hotel matched just 8.9% of the time.
  • Yet 35.6% of hotels formed a steady core that took 88% of all mentions.
  • In each destination, the ten most-mentioned hotels took 57.3% of mentions, rising to 74.1% for luxury queries.

The data was collected in June and July and measures visibility, not bookings. But the lesson is clear. A hotel dropping from first to fourth in one AI check is probably noise. What matters is whether it stays in that recurring shortlist across engines and repeated runs.

Luxury is the hardest shortlist to break into. Brie Moreau ran a live test at the event, asking the room to type “best hotel to stay in Hawaii” into ChatGPT, then asking who hadn’t been given Four Seasons or Ritz-Carlton.

What it means for hotels:

  • Report how often your hotel appears and how often it makes the top three, split by engine and by type of trip (boutique, romantic, luxury).
  • Label every number as measured (GA4, Search Console), modeled (third-party tools) or sampled (AI prompt tracking), as Al Sefati suggested.
  • Ask guests “how did you hear about us?” at booking or check-in.
  • Lead reports with the three to six things the client truly cares about, a tip from Dana DiTomaso of Kick Point. She also suggests giving AI tools a context file with each client’s seasonality and tracking quirks, so a normal seasonal dip isn’t reported as a crisis.

The OTAs Aren’t Waiting

This wasn’t from BrightonSEO, but it’s the context every hotel should keep in mind. PhocusWire reports that Booking Holdings spent $2.37 billion on marketing in Q2 2026 (up 11%), Expedia Group $2.12 billion (up 10%) and Airbnb about $875 million. That’s roughly $5.37 billion in a single quarter.

Expedia CEO Ariane Gorin said the company has been experimenting “in organic with AEO,” with agentic browsers and with paid ads in ChatGPT. PhocusWire also reports she described SEO as stable but soft, while social and answer engine optimization are among Expedia’s fastest-growing channels.

The retail data from BrightonSEO shows where this leads. In Toledano’s case study of laptop brands, visits to the brand’s own site after an AI answer fell from 60.2% across all prompts to 47.5% on purchase-intent prompts, while retailer visits rose from 6.3% to 21.6%. For hotels, the retailer in that story is the OTA.

None of the decks we reviewed showed an AI agent booking a hotel end to end. But Google is already moving AI Mode closer to the booking, so it’s worth watching.

What it means for hotels: you can’t outspend the OTAs, so be more specific than them. Make the case for booking direct on your own pages, with clear perks and best-rate messaging. And protect your branded search. Toledano found 55.9% of AI-influenced visits arrive via search, against 40.4% of other visits, which makes brand defense advertising more important, not less.

What Hotels Should Do Next

Nicolas Rabouille summed it up in his talk title: 95% of GEO is recycled SEO. Benu Aggarwal’s walk-through of Google’s own guidance for AI features made the same point: be crawlable, useful, trusted and consistent. Here’s where we’d start:

  • Run an AI access check. Make sure rates, rooms and amenities are readable without JavaScript, and that pages load fast enough for AI crawlers.
  • Audit what AI gets wrong. Ask the main AI tools about your hotel’s key facts and fix the sources behind any errors.
  • Map who AI cites for you. Then point PR, partnerships and outreach at those sources.
  • Use your guests’ words. Celeste Gonzalez and Martin Canchola both showed how to mine reviews and Reddit threads for page copy, FAQs, Google Business Profile updates and new content, such as a pet policy page built from the questions guests keep asking.
  • Answer the follow-up questions. Pawlowski’s data shows hotel brands vanish from the questions their own searches generate. Publish answers to them, such as which neighborhood to stay in.
  • Plan content by traveler type. Michael Louderback mapped travel personas to content: hotel-led travelers need pages by traveler type, amenity and neighborhood, while families need FAQs and total-cost guidance.
  • Keep content human-led. At Moz, Chima Mmeje’s fully AI-written Learn Center updates gained 23.3% more clicks in two months, then lost 7.75% of traffic the next quarter. Human-led updates to the blog and product pages grew more than 90%.
  • Redirect retired pages properly. Laura Beatty’s advice: a working redirect isn’t enough if it drops people on a generic page. Send expired offers and old pages somewhere that matches what they came for. Her clean-up of thousands of redirects and 404s on one site took non-branded clicks from 36.5K to 72.8K.
  • Measure the shortlist, not the rank. Track appearance rate, branded search and direct bookings together.

Is Your Hotel in the AI Shortlist?

At three&six, we’re helping hotels understand where they appear in AI search, what AI is getting wrong about them and how to earn a place in the answers that lead to bookings.

Book a Discovery Audit to see where your hotel shows up across search, AI and social, or talk to three&six about your hotel’s AI search strategy.

Sources