# AI Will Make Metasearch Obsolete Within a Decade
The Claim
Metasearch — the aggregation of prices from multiple suppliers into a single comparison grid — was built to solve the problem of travel distribution fragmentation. The hypothesis here is that AI is solving the same problem more elegantly: rather than displaying 47 flight options in a grid and asking the traveler to filter and evaluate, a conversational AI can understand the traveler's actual constraints and preferences and deliver a single recommendation that is already optimized. If AI succeeds in this, the metasearch UX model becomes redundant.
The Evidence For
The most striking piece of evidence is structural: Booking Holdings wrote down the value of Kayak — the metasearch platform it acquired for nearly $2 billion — at the same Phocuswright conference where Google announced agentic hotel booking within AI Mode. Wall Street analysts described the write-down as a possible 'early casualty of generative AI's impact on travel planning.' TripAdvisor's organic traffic, meanwhile, is in 'meaningful decline' partly attributed to generative AI absorbing the 'what's the best hotel in Rome' queries that previously flowed through review and metasearch sites.
Sabre's Kurt Ekert described the look-to-book ratio explosion under NDC and agentic AI — where agentic systems will generate 'logarithmically more search queries' — as a specific threat to metasearch economics. If agents are generating vast numbers of comparison queries but completing very few bookings, the cost per conversion for metasearch platforms becomes prohibitive. GBT's Evan Konwiser and Spotnana's Steve Singh both described conversational AI 'already accounting for a meaningful share of bookings' in corporate travel — a segment where metasearch has historically been strong.
The Evidence Against
Skyscanner's Bryan Batista was the most direct industry voice against this hypothesis. He argued — with evident conviction — that metasearch is not being killed by AI; it is being enhanced. His framing of 'metasearch on steroids' reflects a genuine belief that AI tools improve Skyscanner's ability to deliver relevance across 160 million monthly travelers rather than rendering it redundant. He specifically challenged the earlier conference claim that OTAs could become their own search engines, arguing that search is technically hard and that Skyscanner's two decades of investment in that specific domain cannot be quickly replicated.
Kayak's Paul Jacobs made a complementary argument grounded in consumer psychology: at the moment of booking, travelers still want price confirmation and a trusted brand. Even if AI narrows the decision to two or three options, a price-checking confirmation step — what metasearch is fundamentally good at — remains relevant. The Kayak write-down itself may reflect Booking Holdings portfolio rationalization as much as market-driven impairment; treating it as definitive evidence of metasearch's decline involves inferential risk.
Assessment
The hypothesis is directionally plausible but overstated for a ten-year horizon. Metasearch is clearly under structural pressure: its traffic sources are under attack from AI, its unit economics depend on a commission spread that could compress in an AI-native distribution environment, and its 'front door' function in discovery is being replaced by conversational AI for an increasing share of travelers. But the platforms themselves are adapting — Skyscanner is integrating AI tools, and the core consumer need (price assurance at the moment of booking) persists. The most defensible prediction is meaningful metasearch contraction and transformation rather than obsolescence within ten years. Standalone metasearch destinations will struggle; metasearch as an embedded capability within AI-powered booking flows will survive.
**Verdict: Partially supported. Confidence: Medium.**