Booking.com
genAI trip planning assistant
Booking.com has run a conversational AI Trip Planner built on the ChatGPT API since 2023, deployed across five English-speaking markets then extended to the EU and to ChatGPT in October 2025.
Objective
Capture the inspiration and planning phase in natural language to bring travelers to the Booking.com inventory earlier and shorten the path to booking.
The deployment
Booking.com's AI Trip Planner, launched in beta in the United States in June 2023, is a conversational assistant built on in-house ML models and OpenAI's ChatGPT API. The traveler asks open-ended questions (where to go for a romantic weekend), receives destination and accommodation suggestions with prices and booking links, and refines within the conversation. The suite expanded in 2024 with Smart Filter (natural language search), Property Q&A (answers drawn from listings), and Review Summaries. In October 2025, Booking.com also became one of the first partners for apps inside ChatGPT via OpenAI's Apps SDK.
Results Proof C
Official Booking.com releases documenting the multi-market rollout and the feature upgrades, picked up by the specialist press (Skift). Evidence of scale = multi-market rollout and more than 12 months in production; the public performance metrics remain qualitative.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- llm OpenAI ChatGPT API LLM API underlying the AI Trip Planner; the exact model version is not specified in the sources.
- outil Modeles ML maison Booking.com Proprietary recommendation and personalization layers combined with the ChatGPT API, anchored to live inventory and prices.
- plateforme OpenAI Apps SDK SDK that carries the Booking.com app inside ChatGPT since October 2025, among the first partners.
How it runs, concretely
For ops teams-
1Open-ended question client
The traveler makes an inspiration or planning request in natural language.
-
2Suggestion generation AI
The assistant combines in-house ML and the ChatGPT API to propose destinations and accommodations.
-
3Anchoring to the inventory AI / data team
It pulls prices, availability, and booking links from the Booking.com database.
-
4Refinement and handoff client
The traveler refines within the conversation then moves to the listing and the booking.
The intent expressed in natural language, crossed with Booking.com's live inventory and prices. If the assistant does not surface real availability, it remains an inspiration engine with no transactional value.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- inventory and prices accessible via API
- rich product listings for the Q&A
- customer reviews for the summaries
Org prerequisites
- AI product team
- governance of multi-market personalization
- country-by-country rollout process
Possible stack
- commercial LLM API
- RAG layer on the inventory and listings
- existing search engine
The plan, step by step
- Step 1Connect a natural language search to the existing search engine, on a single market, in A/B test against the classic filters.Deliverable: Natural language search in test with a measurement plan.
- Step 2Build the conversational assistant backed by live inventory and prices via API, with real booking links.Deliverable: Assistant in beta, anchored to real availability.
- Step 3Add the product Q&A drawn from the listings and the review summaries to cover the reassurance phase.Deliverable: Q&A features and review summaries in production.
- Step 4Measure engagement, customer service contacts, and conversion against the classic journey, then launch the rollout country by country.Deliverable: Per-market assessment and expansion plan.
First step: Start with natural language search (Smart Filter) on one market, before adding planning and product Q&A.
The same pattern, proven in other industries
Amazon
genAI shopping assistant
Burberry
genAI copilot for consultants (search, image recognition, complete-the-look)
Flipkart
conversational genAI shopping assistant
Sources
- S1 Booking.com Launches New AI Trip Planner to Enhance Travel Planning Experience Primary archive pending
- S2 Booking.com Enhances Travel Planning with New AI-Powered Features for Easier, Smarter Decisions Primary archive pending
- S3 ChatGPT Travel Apps Are Now a Thing - Starting with Booking.com and Expedia Established press archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.