Gemini Overtook Perplexity in AI Referral Traffic. What Travel Brands Should Do Now

Travel brands should stop optimizing for ChatGPT and Perplexity alone.

That is the main takeaway from the latest referral traffic shift. New March 2026 reporting says ChatGPT still leads AI chatbot referrals at 78.16%, but Gemini rose to 8.65% and passed Perplexity at 7.07%. For hotels, tour operators, DMCs, and resort groups, that is not a trivia point. It is a distribution warning. If your visibility strategy only covers Google SEO plus a little ChatGPT experimentation, you are already behind the way trip planning is fragmenting. Source: Marketing Edge reporting on March 2026 referral shares.

The practical move is simple. Treat AI discovery as a multi-engine channel. Build content and structured data that can be cited across ChatGPT, Gemini, Perplexity, and Google AI products at the same time. The brands that do this early should gain direct bookings, reduce OTA dependence, and become easier for AI assistants to recommend. For hotel operators, the clearest commercial starting point is AI Visibility for Hotels, followed by the Travel AI Audit to see which engines still miss the brand.

That matters because the commission math is brutal. Hotels still give up roughly 15% to 25% per OTA booking, depending on platform, market, and program participation. Palmtree.ai exists for exactly this problem: make travel businesses visible to AI engines so more travelers find and book them directly instead of discovering them only inside OTA ecosystems.

Why this traffic shift matters more in travel than in most industries

Travel discovery is unusually vulnerable to interface changes.

A traveler does not just search one keyword and buy. They compare neighborhoods, seasons, transportation, reviews, cancellation policies, family fit, price bands, and itinerary logic. That is exactly the kind of messy decision process where AI assistants are useful. They compress research, summarize options, and turn ten browser tabs into one recommendation flow.

That makes AI referral share strategically important. If Gemini is growing faster than Perplexity, travel brands need to understand what that means operationally.

1. Discovery is splitting across engines

Search behavior is no longer concentrated in one place. Search Engine Journal reported that Google’s global search share fell to 90.01% in March 2026, while AI search usage keeps expanding outside classic search measurement frameworks. Source: Search Engine Journal.

Google still dominates, but dominance is not the same thing as monopoly over discovery behavior. The real shift is that high-intent research is now distributed across search, chat, map products, and assistants.

For travel brands, that means one broken assumption needs to die: ranking in Google does not guarantee recommendation visibility inside AI planning flows.

2. Gemini growth changes the optimization mix

Perplexity has earned a lot of attention because its interface is citation-heavy and easy for marketers to study. But Gemini sits much closer to Google’s ecosystem, which makes it especially relevant for travel.

Hotels already depend on Google surfaces for maps, reviews, local intent, directions, and business profile signals. If Gemini becomes a larger discovery surface, those local authority assets become even more important.

In plain English, the travel brand with strong entity clarity, complete structured data, real review density, and destination-level content now has more paths into recommendation systems.

3. OTAs are better positioned than most property websites

This is the uncomfortable part.

OTAs have structured inventory, review volume, massive brand authority, and standardized listing data. AI systems like clean, comparable, frequently updated information. OTAs have a lot of it.

That does not mean hotels should leave Booking.com or Expedia. It means they should stop treating OTA participation and direct booking growth as opposites. The better strategy is OTA-smart, not anti-OTA.

That is also the strategic position Palmtree.ai takes. Stay visible where AI engines already pull trust signals, but build enough AI-native authority that your direct site starts appearing earlier in the consideration journey.

The data points travel marketers should care about right now

Here are the numbers worth anchoring on this week.

One caution: marketers love acting like every percentage change is a revolution. It is not. Gemini passing Perplexity does not mean Perplexity stops mattering. It means your workflow needs broader coverage.

What hotels and tour operators should change this week

The best response is not “make more AI content.” That usually produces thin pages, generic articles, and zero recommendation lift.

The right response is to improve the assets AI systems actually consume and trust.

1. Fix your entity clarity across the open web

If an AI model cannot confidently understand who you are, where you are, what you sell, and why you are relevant, you will not be recommended consistently.

For hotels, that means your property identity needs to match across:

• website title tags and on-page copy
• Google Business Profile
• schema markup
• booking and review platforms
• social bios
• major destination listings
• press mentions and local partnerships

This sounds basic, but it is where a lot of travel brands quietly break. The hotel uses one name on Google, a slightly different brand variant on Booking.com, a weak homepage title, and a vague about page. AI systems then have to guess whether all those records describe the same entity.

Start there.

If you need a technical baseline, the strongest next read is our guide on schema markup for hotels.

2. Build pages that answer planning questions, not just booking questions

Most hotel sites are optimized for someone who already knows the property.

That is too late.

AI recommendation flows often start with prompts like:

• best boutique hotel in Lisbon for a long weekend
• family-friendly beach resort in Puglia with easy airport transfer
• boutique hotel in Rome near great food but not too touristy
• luxury safari lodge with direct transfers and strong reviews

If your site only has room pages and a generic amenities page, it is missing the layers that answer those prompts.

Travel brands need more destination-intent content, including:

• neighborhood guides
• stay-type pages for couples, families, remote workers, food travelers, wellness travelers
• best-time-to-visit pages
• itinerary pages
• local comparison pages
• FAQ pages built from real booking questions

This is where many AI visibility programs fail. They create “top 10 travel trends” content instead of pages that make recommendation engines more confident about matching the property to traveler intent.

3. Treat reviews as retrieval assets

Reviews are not only persuasion tools anymore. They are machine-readable relevance signals.

If a traveler asks an AI assistant for a quiet family-run hotel with excellent breakfast and walkability, the assistant may infer those traits from review language, review summaries, local citations, and owner responses.

So the goal is not only “get more reviews.” The goal is:

• get detailed reviews mentioning concrete experiences
• respond with useful specifics, not canned thanks
• reflect recurring strengths in your website copy and FAQ
• close the loop between guest language and structured content

That is part of why review response quality matters. According to Tripadvisor research cited in Palmtree’s strategic positioning, travelers are 77% more likely to book when owners respond to reviews. In the AI era, those responses also create useful text for retrieval and summarization.

4. Upgrade your structured data beyond the bare minimum

A lot of hotel schema implementations are technically present but strategically weak.

Yes, you need LodgingBusiness, Hotel, or related schema types. But you also need enough specificity to help systems distinguish your property.

High-value fields include:

• amenity detail
• check-in and checkout policies
• room types
• family or pet suitability
• parking and transport availability
• nearby landmarks
• aggregate review data
• FAQ schema where appropriate
• image metadata and consistent media naming

You can go deeper in our article on how to get your hotel recommended by AI, but the short version is this: basic schema gets you parsed, richer schema gets you matched.

5. Stop thinking channel by channel

Travel marketers are used to channel silos. SEO here, social there, OTA content somewhere else, review ops in another dashboard.

AI discovery punishes that setup.

The assistant does not care which internal team owns the truth. It cares whether the truth is consistent across sources.

Your website says adults-only. Your OTA listing says family friendly. Your Instagram bio says wellness retreat. Your reviews complain about noise. Your FAQ does not address transport. That inconsistency lowers recommendation confidence.

The fix is not a bigger content calendar. It is a better source-of-truth system.

Why Gemini’s rise may favor disciplined brands

There is a useful contrarian angle here.

When a new AI discovery surface becomes important, many brands respond with volume. They publish a flood of generic AI-written pages. That may increase page count, but it rarely improves recommendation quality.

Gemini’s rise may actually reward the opposite behavior: cleaner entities, stronger Google ecosystem signals, better local authority, better page structure, and more trustworthy data.

That is good news for disciplined independent brands. Large OTAs still have scale, but smaller hotel groups and operators can move faster when they stop wasting effort on shallow content.

A practical 30-day playbook for travel brands

Here is the operating plan I would use if I ran growth for an independent hotel group, DMC, or tour operator right now. If you want the supporting trust layer too, pair it with Palmtree’s Travel AI Score Methodology and the Hotel AI Visibility Benchmark 2026.

Week 1: Audit the recommendation layer

Check how your brand appears across:

• ChatGPT
• Gemini
• Perplexity
• Google Business Profile
• top OTA listings
• Tripadvisor and key review sites

Look for factual gaps, missing differentiators, wrong positioning, and missing local context.

Week 2: Fix trust and structure

Prioritize:

• homepage positioning clarity
• schema completeness
• review response quality
• consistent naming and entity details
• stronger FAQ coverage
• clearer destination and traveler-fit pages

Week 3: Publish intent-matching content

Create the pages most likely to support recommendation prompts:

• who this property is best for
• what nearby experience clusters matter
• why book direct with us
• seasonal local guides
• detailed planning FAQ

Week 4: Measure recommendation lift, not vanity metrics

Track:

• AI engine mentions
• citation frequency
• referral sessions from AI tools
• direct booking inquiries influenced by AI
• branded search lift after publication
• OTA share vs direct share over time

This is basically the Palmtree.ai model in practice: measure AI travel visibility, fix the data layer, publish high-intent content, and convert discovery into more direct demand.

What tour operators and DMCs should do differently

Hotels are not the only winners here.

Tour operators and DMCs may benefit even faster because AI trip planning often needs itinerary logic, niche expertise, and confidence in local execution.

That means operators should emphasize:

• destination expertise pages
• sample itineraries
• specialty trip formats
• local guide credibility
• operational clarity around transfers, group sizes, and inclusions
• proof of trust through testimonials and review language

If you run tours or destination services, read our broader playbook on the tour operator and DMC AI shift.

The strategic mistake to avoid

Do not turn this into a Perplexity-versus-Gemini debate.

That is the wrong frame.

The real issue is whether your travel brand has become easy for multiple AI systems to understand, trust, and recommend.

Perplexity still matters. ChatGPT matters most. Gemini matters more than it did a month ago. Google still matters massively. And the next interface shift will happen before most travel brands finish fixing their homepage copy.

So optimize the underlying recommendation assets, not just the current leaderboard.

FAQ

Does Gemini passing Perplexity mean Perplexity no longer matters for travel brands?

No. Perplexity still matters, especially because its citation-first interface makes it a useful environment for testing how your brand is being sourced. The important change is that Gemini now deserves equal or greater attention in many travel workflows.

Why is Gemini especially relevant for hotels?

Because Gemini sits close to Google’s ecosystem, where hotel discovery already depends on maps, local business data, reviews, and structured property information. Brands with stronger Google-facing data quality may have an easier path into Gemini-led discovery.

Should hotels leave OTAs if AI discovery is growing?

No. That is a bad strategy for most properties. OTAs still provide demand, reviews, and authority signals. The smarter move is to stay OTA-visible while improving AI and direct-booking visibility so your dependency drops over time.

What is the fastest technical fix for AI visibility?

Usually schema completeness plus entity consistency. If your property details, amenities, policies, and review signals are inconsistent across the web, fix that before publishing more content.

What kind of content helps most with AI travel recommendations?

Pages that answer real traveler intent. Neighborhood guides, who-this-is-for pages, local logistics pages, direct booking benefit pages, itinerary pages, and detailed FAQs usually help more than generic trend content.

How should travel brands measure success here?

Track AI mentions, citations, referral traffic from AI tools, direct inquiry lift, and the share of bookings or leads that arrive after AI-assisted discovery. Pure pageview metrics are not enough.

Final take

Gemini overtaking Perplexity is not the whole story. It is the signal.

The signal is that AI trip planning is spreading across more surfaces, and travel brands that still treat AI visibility as a side project will lose recommendation share to OTAs and better-prepared competitors.

The brands that win this year will not be the ones publishing the most content. They will be the ones with the clearest entities, strongest review and structured-data layer, and the best answers to traveler intent.

That is the work Palmtree.ai is built to do.