# How to Measure AI-Assisted Bookings Without Calling Every AI Click Direct A traveller asks an AI assistant for a hotel, sees your property, compares prices on an OTA and finally books on your website. Which channel deserves credit? The answer depends on what you are measuring. Discovery, the completed transaction and the long-term guest relationship can belong to different channels. This distinction matters more as booking flows change. Google's AI Mode hotel experience includes both hotel groups and online travel sites as partners, while restaurant assistants can show availability from reservation platforms. [Google's hotel announcement](https://blog.google/products-and-platforms/products/search/book-travel-ai-mode/) and [OpenAI's restaurant booking guide](https://help.openai.com/en/articles/9237897-searching-the-web-with-chatgpt) make clear that an AI-assisted choice is not necessarily a direct booking. ## Record three separate events **Discovery:** Where did the guest first notice or research the business? This might be an AI answer, Google Maps, an OTA, a destination guide or a recommendation from a friend. It is often hard to know with certainty; an optional “How did you hear about us?” question can add useful context but will not produce perfect attribution. **Reservation:** Where was the booking completed, and who handled payment and customer service? This determines whether to classify the transaction as a hotel-site booking, an OTA reservation, a restaurant marketplace booking or another route. **Relationship:** Did the guest later opt into direct communication, join a program and return? A guest introduced by an intermediary may become a valuable direct repeat guest. Measure that change without relabeling the first booking. ## Build a small channel report For each source and booking path, track confirmed bookings, cancellations, attended stays or covers, gross value, platform fees, and estimated contribution after acquisition cost. Add repeat bookings over a consistent period where guest consent and data quality allow it. Report unknown sources honestly. Hotels using Google free booking links can review Hotel Center performance and add tracking to their booking pages. [Google explains reporting and click tracking](https://support.google.com/hotelprices/answer/10472393?hl=en). Restaurants should compare their website reservations with reports from each booking provider; the [reservation systems guide](https://loop.fans/blog/restaurant-reservation-systems-guide) explains which reports and guest data to ask for. Keep the definitions beside the numbers so a dashboard does not quietly mix clicks, reservation requests and completed visits. ## Treat AI visibility as a diagnostic Check whether assistants describe your business accurately and include a working action link. A mention can reveal an information problem or an opportunity, but it is not a sale. Results vary by prompt, location and time. If you run a restaurant or bar, Loop's [AI Visibility Checker](https://loop.fans/free-tools/ai-visibility-checker) offers a starting view of AI recommendations. Hotels can test realistic traveller prompts directly. Compare either kind of visibility finding with actual booking outcomes. For hotels, the [direct booking link guide](https://loop.fans/blog/google-hotel-direct-booking-links-ai) covers the transaction path. For restaurants, see the [AI reservation guide](https://loop.fans/blog/restaurant-ai-reservations-chatgpt-google). [Loop's hotel](https://loop.fans/hotels) and [hospitality](https://loop.fans/hospitality) programs can help build the ongoing relationship after a first booking.