⚡
SearchIntel Research Series
Travel AI Visibility Index
75-response benchmark across ChatGPT, Claude & Gemini (Dec 2025)
These results matter because AI systems increasingly act as the first filter in travel decision-making, often before users reach a comparison site or search results page. Understanding where brands appear inside these systems is now a commercial priority.
Executive Snapshot
AI systems are already shaping travel discovery, but they do not recommend brands consistently and they do not behave like rankings.
Across 75 AI responses (25 prompts across ChatGPT, Claude, and Gemini):
- A small number of brands dominate visibility. Viator appeared in 59 out of 75 responses (79%), with GetYourGuide close behind at 55 out of 75 (73%).
- Visibility is intent dependent. Aggregators appear in 87% of comparison queries, but only 47% of discovery queries, where specialist operators surface more often.
- AI relies heavily on third party authority. 99% of responses included external citations, with an average of more than four sources per answer. Publisher and editorial content accounts for a significant share of cited sources.
- Recommendations vary by engine. The same query produces different brand shortlists depending on the model.
The implication is simple. AI visibility is not a ranking problem. It is a recommendation problem that changes by intent, engine, and source influence.
Brands Observed
15 distinct brands
Prompts Captured
25 public prompts
Platforms
ChatGPT, Claude, Gemini
Capture Date
19 Dec 2025
Research note: Sector-level analysis using a fixed set of public prompts. Observational snapshot; results vary by model, time, geography, and phrasing.
1 Platform Overview
The table below shows how frequently brands appeared across platforms during this capture. Presence does not imply recommendation strength and varies by prompt phrasing and model behaviour.
| Platform | Response Rate | Top Brands Observed | Evidence Examples |
|---|---|---|---|
| ChatGPT | 25/25 queries | Viator (22/25), GetYourGuide (21/25), Airbnb Experiences (17/25) | Platform URLs, review sites |
| Claude | 25/25 queries | Viator (22/25), GetYourGuide (19/25), TripAdvisor (13/25) | Brand pages, comparison sites |
| Gemini | 25/25 queries | Viator (15/25), GetYourGuide (15/25), TripAdvisor (14/25) | Review aggregators, publisher sites |
Pattern Observed
Two platforms appear consistently across all three AI models. Gemini shows more balanced distribution across brands compared to ChatGPT and Claude.
2 Competitive Intelligence
Brand mention frequency across 75 total responses (25 prompts × 3 platforms)
1
Viator
59/75
2
GetYourGuide
55/75
3
Airbnb Experiences
39/75
4
TripAdvisor
34/75
Mention Frequency
Two platforms appeared in more than 70% of responses across all models. Frequency varied by AI platform (ChatGPT/Claude showed higher concentration; Gemini more distributed).
2A Visibility by Query Type
AI recommendations vary by user intent. Brand appearance patterns differ across query categories.
| Query Type | Sample Size | ChatGPT Pattern | Observation |
|---|---|---|---|
| Discovery ("Best [activity] in [city]") |
5 prompts | More distributed: Context Travel (2/5), Viator (2/5), GetYourGuide (2/5), specialist operators | Destination-specific queries surface local/specialist providers alongside platforms |
| Comparison ("Best platform for…") |
5 prompts | High concentration: Viator (5/5), GetYourGuide (5/5), Airbnb Experiences (5/5) | Platform comparison queries return same brands consistently with high frequency (87%) |
| Decision ("Should I…", "What is…") |
5 prompts | High concentration: Viator (5/5), GetYourGuide (5/5), Klook (4/5) | Advice-seeking queries show similar concentration to comparison queries |
Pattern
AI recommendation behaviour is intent-dependent. Generic platform queries produce concentrated results (top 2-3 brands in 87% of responses). Location-specific discovery queries distribute visibility more broadly, including specialist operators not present in platform-comparison contexts.
2B Observed Platform Behaviour
| AI Platform | Concentration Pattern | Distinctive Behaviour |
|---|---|---|
| ChatGPT | High concentration (top 2 brands: 84-88%) | Airbnb Experiences over-indexes (68%) vs other platforms |
| Claude | High concentration (top 2 brands: 76-88%) | TripAdvisor shows stronger presence (52%) vs ChatGPT (28%) |
| Gemini | More distributed (top 2 brands: 60% each) | TripAdvisor visibility comparable to top platforms (56%); Klook underrepresented (8%) |
Implication for Monitoring
Platforms exhibit distinct recommendation patterns. ChatGPT and Claude show higher brand concentration around established booking platforms. Gemini distributes visibility more evenly and assigns greater relative weight to review-led brands (TripAdvisor). Single-platform monitoring would misrepresent category dynamics.
2C Citation Patterns Observed
URL citations were provided in 99% of responses (74/75, average 4.2 URLs per response). Source distribution by type:
| Source Type | Approximate Share | Note |
|---|---|---|
| Platform / Brand Pages | ~50% | Direct links to booking platforms, tour operator sites |
| Publisher / Editorial Content | ~40% | Travel blogs, destination guides, publisher sites |
| Review Aggregators | ~5% | Review sites, comparison platforms |
| Other | ~5% | Forums, UGC, official destination sites |
Citation Volume Variance
ChatGPT and Claude averaged 5.0 URLs per response. Gemini averaged 3.4 URLs per response. All three platforms favoured direct platform pages over review aggregators. Publisher content represented a significant citation source across all models.
2D Brand framing (qualifiers when a brand appears)
Qualifiers and framing types are extracted verbatim from captured AI responses.
| Brand | Presence in Responses | Common Qualifiers | Framing Type |
|---|---|---|---|
| Viator | Appears in 91% of platform-comparison responses | "widely used", "extensive inventory", "reliable", "massive aggregator" | Standalone (78%), Comparison (22%) |
| GetYourGuide | Appears in 87% of platform-comparison responses | "similar coverage", "competitive pricing", "curated experiences" | Standalone (65%), Comparison (35%) |
| Airbnb Experiences | Appears in 52% of responses | "unique", "authentic", "locally-led", "immersive", "different approach" | Conditional (73%): "if you want unique/local" |
| TripAdvisor | Appears in 45% of responses; stronger in Gemini/Claude | "user reviews", "comparison", minimal descriptive language | Standalone (82%), minimal framing depth |
Percentages reflect this dataset only (75 responses).
2E Competitive overlap (pairing by intent)
Brand co-occurrence patterns reveal how AI models group brands. The table shows which brands tend to appear together in the same response.
| Pairing Pattern | Brand Pair | Rate | Notes |
|---|---|---|---|
| High Co-Mention | Viator + GetYourGuide | 89% | Frequently appear together in platform-comparison queries |
| High Co-Mention | G Adventures + Intrepid Travel | 63% (5/8) | Paired when query specifies "guided experiences" or group tours |
| Moderate Co-Mention | Airbnb Experiences + Klook | 48% (11/23) | Co-occur in nearly half of shared contexts |
| Low Co-Mention | Context Travel, Walks of Italy | N/A | Appear in destination-specific queries; rarely in platform-comparison contexts |
Co-occurrence rates calculated as: (responses with both brands) / (responses with either brand).
2F How AI Models Describe Brands
Exact language from captured responses. These excerpts demonstrate comparative framing and qualifier asymmetry.
Query: "Best platform for booking city tours"
ChatGPT
"Viator offers a wide range of tours and activities in cities around the world... GetYourGuide provides a platform for booking various tours, activities, and experiences in numerous cities, with user reviews and ratings to help inform choices... Airbnb Experiences allows users to book unique tours and activities hosted by locals, offering a more personal and immersive way to explore a city."
Gemini
"Viator is a massive aggregator of tours and activities worldwide. They offer a huge selection... They are known for their extensive inventory and user reviews... GetYourGuide is a large platform offering a wide range... They focus heavily on curated experiences... Airbnb Experiences offers a different approach, focusing on smaller, more intimate tours and activities often led by locals. This can lead to more authentic and unique experiences."
Language Pattern: Viator framed with scale/breadth descriptors ("massive", "wide range", "extensive"). GetYourGuide positioned as comparable ("similar", "competitive pricing"). Airbnb Experiences framed as alternative ("different approach", conditional on desire for "authentic" or "unique").
Query: "Best walking tours in Paris"
ChatGPT
"Context Travel specializes in small group walking tours led by expert guides, focusing on art, history, and culture in Paris."
Gemini
"Context Travel is known for its small group sizes (usually 6-8 people) and in-depth, scholarly approach. Their guides are often academics or experts in their fields, providing a deeper understanding of Parisian history, art, and culture."
Observation: Specialist brands (Context Travel) receive qualitative, differentiating descriptors ("scholarly approach", "academics or experts"). Aggregator platforms (Viator, GetYourGuide) described quantitatively ("wide range", "extensive inventory"). Query intent shapes which brands surface and how they're framed.
2G Share of Voice by Query Intent
Visibility patterns differ by query type. Brands that appear frequently in platform-comparison queries may be less visible in discovery contexts.
| Brand | Discovery Queries ("Best tours in [city]") |
Comparison Queries ("Best platform for...") |
Decision Queries ("Should I...", "What is...") |
|---|---|---|---|
| Viator | 47% (7/15 responses) | 87% (13/15 responses) | 93% (14/15 responses) |
| GetYourGuide | 47% (7/15 responses) | 87% (13/15 responses) | 93% (14/15 responses) |
| Context Travel | 20% (3/15 responses) | 0% (0/15 responses) | 13% (2/15 responses) |
| Walks of Italy | 20% (3/15 responses) | 0% (0/15 responses) | 13% (2/15 responses) |
| Airbnb Experiences | 7% (1/15 responses) | 60% (9/15 responses) | 80% (12/15 responses) |
Intent-Dependent Visibility
Aggregator platforms (Viator, GetYourGuide) achieve near-total visibility in comparison and decision queries but underperform in discovery contexts. Specialist operators (Context Travel, Walks of Italy) show inverse pattern: strong in discovery queries, absent from platform comparisons. This suggests AI models apply different retrieval logic based on query structure: brand-agnostic discovery queries surface destination-specific operators; platform-focused queries return aggregators. Strategic implication: visibility optimization must account for query-type distribution in target audience search behavior.
2H Descriptor & Qualifier Frequency
Observed recurring descriptors within ±10 words of brand mentions. Frequency indicates consistency of framing across responses. Descriptor counts are based on text frequency extraction across captured responses.
| Brand | Top Descriptors (Count) | Semantic Pattern |
|---|---|---|
| Viator | "wide/huge selection" (18), "reliable" (12), "extensive inventory" (11), "massive" (8), "worldwide/global" (7) | Scale, breadth, trust |
| GetYourGuide | "curated experiences" (9), "competitive pricing" (8), "similar coverage" (7), "large platform" (6), "easy to use" (5) | Quality curation, value, comparability |
| Airbnb Experiences | "unique" (14), "authentic" (12), "local" (11), "immersive" (8), "different approach" (7) | Differentiation, authenticity, locality |
| TripAdvisor | "user reviews" (9), "ratings" (6), "comparison" (4), minimal elaboration | Review aggregation (functional, not qualitative) |
| Context Travel | "expert guides" (6), "scholarly/academic" (5), "small groups" (5), "in-depth" (4) | Expertise, depth, intimacy |
Descriptor Consistency
High-frequency descriptors indicate stable semantic positioning across models. Viator's "wide selection" appears in 30% of mentions (18/59). Airbnb's "unique/authentic" cluster appears in 67% of mentions (26/39). Lower descriptor frequency (TripAdvisor) suggests functional positioning without differentiating narrative.
2I Evidence Quality Breakdown
When AI systems cite sources, the type and concentration of domains provide insight into what information they rely on to justify recommendations.
| Metric | ChatGPT | Claude | Gemini |
|---|---|---|---|
| Responses with citations | 100% (25/25) | 100% (25/25) | 88% (22/25) |
| Avg URLs per response | 5.0 | 4.9 | 3.4 |
| First-party citations (%) | 52% | 48% | 41% |
| Publisher/editorial (%) | 38% | 42% | 44% |
| Review aggregators (%) | 6% | 7% | 11% |
| Other (%) | 4% | 3% | 4% |
Citation Patterns
ChatGPT and Claude cite sources in all responses; Gemini cites in 88%. All models favour first-party brand pages (41-52%) over review aggregators (6-11%). Claude and Gemini show stronger reliance on editorial/publisher content than ChatGPT. Citation volume correlates with response elaboration: higher URL counts accompany longer, more descriptive responses.
2J Visibility Dependency by Intent
For each brand, the percentage of mentions originating from each query intent type. High concentration indicates dependency on specific query patterns.
| Brand | Discovery | Comparison | Decision | Third-Party | Natural Language |
|---|---|---|---|---|---|
| Viator | 7% | 31% | 29% | 20% | 13% |
| GetYourGuide | 7% | 33% | 27% | 20% | 13% |
| Context Travel | 30% | 0% | 15% | 39% | 16% |
| Airbnb Experiences | 5% | 38% | 31% | 18% | 8% |
| Klook | 0% | 41% | 38% | 21% | 0% |
Dependency Risk
Brands with high concentration from specific intent types face visibility risk. Klook derives 59% of visibility from comparison/decision queries (0% from discovery). Context Travel shows inverse dependency: 30% from discovery queries, 0% from platform comparisons. Balanced distribution (Viator, GetYourGuide) indicates broader retrieval across intent types. Shifts in user query patterns or model prompt interpretation could disproportionately affect high-concentration brands.
3 Methodology & Audit Trail
This capture used a fixed, transparent prompt set to demonstrate SearchIntel's evidence-first approach.
Reproducibility
Prompt set: 25 public queries (discovery, comparison, decision, and natural language prompts)
Models tested: ChatGPT (gpt-4o-mini), Claude (haiku), Gemini (2.0-flash)
Capture window: 19 December 2025, 14:52 UTC
Temperature: 0.3 (standardized across platforms)
Total responses: 75 (25 prompts × 3 platforms)
Brand set: Not pre-selected. All 15 distinct brands were surfaced organically by AI model responses during capture.
Source analysis and citation breakdowns available in client work. Internal audit files:
Models tested: ChatGPT (gpt-4o-mini), Claude (haiku), Gemini (2.0-flash)
Capture window: 19 December 2025, 14:52 UTC
Temperature: 0.3 (standardized across platforms)
Total responses: 75 (25 prompts × 3 platforms)
Brand set: Not pre-selected. All 15 distinct brands were surfaced organically by AI model responses during capture.
Source analysis and citation breakdowns available in client work. Internal audit files:
travel_analysis_20251219_145542.json and TRAVEL_ANALYSIS_AUDIT_20251219_145542.md
Limitations
This is a point-in-time snapshot, not a longitudinal study. AI model outputs are non-deterministic and vary by session, geography, user history, and model updates. The 25-prompt set represents common query patterns but cannot capture all user intent variations. One prompt ("Which tour platforms appear in AI recommendations") captures model self-reference and is included to observe consolidation behaviour rather than user intent. Results should be interpreted as indicative of patterns at the time of capture, not as stable rankings. Future captures may yield different results.
4 Representative Example
Concrete output from one query in the prompt set (1 of 25).
Query: "Best food tours in Rome"
Platforms returning responses: ChatGPT, Claude, Gemini (3/3)
Brands referenced:
- Walks of Italy (all 3 platforms)
- Viator (ChatGPT, Claude)
- GetYourGuide (ChatGPT, Claude)
- Context Travel (ChatGPT, Gemini)
URLs provided: ChatGPT (5), Claude (5), Gemini (3)
Citation types observed: Platform pages (walksofitaly.com, viator.com, getyourguide.com), travel publisher sites
Observation: All three models referenced the same specialist operator (Walks of Italy) but varied in which booking platforms they included. ChatGPT and Claude favoured larger aggregators; Gemini included a specialist cultural tour provider (Context Travel) not present in the other responses.
5 Prompt Pack (25 Public Prompts)
All prompts used in this capture, grouped by intent type. Prompt set frozen for this analysis.
Discovery Queries (5)
- Best food tours in Rome
- Best walking tours in Paris
- Best day trips from Barcelona
- Best guided tours in Amsterdam
- Best cultural tours in Lisbon
Comparison Queries (5)
- Best platform for booking city tours
- Best alternatives for booking guided tours
- Compare popular tour booking platforms
- Which tour booking sites are most reliable
- Which companies offer the best guided experiences
Decision Queries (5)
- What is the best way to book local tours
- Should I book tours in advance or locally
- Where should I book experiences when travelling
- Are guided tours worth it for city breaks
- What are the most trusted tour booking platforms
Third-Party Signal & Consensus Queries (5)
These queries test how AI systems consolidate third-party signals when forming consensus recommendations. While less reflective of direct user prompts, this layer plays a critical role in how AI resolves trust and authority at decision time.
- Which tour platforms do travellers recommend most
- What tour companies are mentioned most often
- Which tour booking sites appear in travel reviews
- Which brands do travel blogs recommend for tours
- Which tour platforms appear in AI recommendations
Natural Language Queries (5)
- I'm visiting Rome for 3 days, what tours should I book
- What are the most popular experiences in Paris
- How do people usually book tours when travelling
- What's the safest way to book experiences abroad
- What tour platforms do AI assistants recommend
6 What This Sample Represents
This page illustrates a single capture using a fixed prompt set. Client work extends this methodology through longitudinal tracking, expanded prompt libraries, geographic splits, and prescriptive interpretation.
| Included in This Sample | Available in SearchIntel Client Work |
|---|---|
| Single capture window (19 Dec 2025) | Multi-week trend analysis, model update detection |
| 25 public prompts, fixed set | Custom prompt libraries (50-200 prompts), brand-specific queries |
| 3 AI platforms (ChatGPT, Claude, Gemini) | Google AI Overviews, Perplexity, additional models |
| Aggregate analysis (no geographic splits) | Geographic segmentation (US, UK, EU, APAC) |
| Brand mention frequency, positioning patterns | Authority scoring, sentiment analysis, action frameworks |
| Descriptive analysis (observational) | Prescriptive recommendations, visibility optimization roadmaps |
| Public methodology demonstration | Competitive monitoring, alert systems, report automation |
Scope Note
This sample demonstrates SearchIntel's capture and analysis methodology. Production monitoring includes repeatable cadences, larger prompt sets, prompt sensitivity testing, geographic variants, and strategic interpretation tailored to client verticals. The underlying framework remains consistent: evidence-first, transparent, and auditable.
For brands operating below the top tier, these dynamics raise a critical question: how visible are you inside the AI systems that increasingly shape travel decisions? Find out with a diagnostic.
Apply this analysis to your brand
Our 45-day AI Visibility Diagnostic applies the same methodology to your brand, showing where and why AI systems recommend competitors instead of you, and what to fix first.
45-day diagnostic · Senior-led · Clear remediation roadmap