Fliggy Launches 'Fliggy Bangbang': AI Moves into the Transaction Phase

On August 10, Fliggy launched its travel AI product, "Fliggy Bangbang."

Instead of adopting a standalone AI chat box interface, this product is directly embedded into Fliggy's existing pages for search, itineraries, orders, as well as flights, hotels, and homestays. It aims to push AI beyond itinerary generation and informational Q&A further into the transaction and fulfillment stages.

According to Fliggy, users can state specific needs in natural language, and Fliggy Bangbang will handle itinerary planning, flight and hotel screening, as well as tasks such as cancellations and rebooking, check-in and seat selection, hotel room upgrades, airport transfer bookings, and invoicing for reimbursement. After the trip, the system can also generate travel memory content.

This means that Fliggy is doing more than just creating an AI search box. In the past, the core workflow of Online Travel Agencies (OTAs) involved search, price comparison, and booking, requiring users to separately manage demands for flights, hotels, and airport transfers. Fliggy Bangbang attempts to let AI directly leverage the platform's existing products, orders, and service capabilities, linking multiple steps together.

Zhuang Zhuoran, CEO of Fliggy, offered a clearer judgment on this shift.

He stated that AI will change the service logic of OTAs in the mobile internet era. In the future, there may no longer be "Online Travel Agencies," but rather "Omni-intelligence Travel Agents."

Zhuang believes that the viability of travel AI does not hinge on its ability to generate a seemingly complete itinerary. "In the travel sector, whether AI can truly get things done depends on whether it can access real supply, cross into real transaction systems, and handle unexpected events in the physical world."

This judgment points to the core of OTA business.

Travel products heavily rely on real-time inventory, pricing, cancellation and change rules, and offline fulfillment. Situations such as flight delays and hotel changes directly impact subsequent itineraries. If AI is only responsible for content generation, the technical barrier is relatively low; however, once it enters order processing, after-sales service, and fulfillment, it relies on the supply chain and transaction systems accumulated by the platform over many years.

Zhuang revealed that Fliggy is currently advancing its AI layout around "3 Es": AI for Experience, AI for Ecosystem, and AI for Enterprise, corresponding to enhancing user experience, prospering the travel ecosystem, and reshaping organizational structure, respectively. Fliggy Bangbang falls under the component focused on user experience.

The product is still in a phase of rapid iteration. Zhuang introduced that Fliggy Bangbang is currently updated on a weekly basis, with features such as multi-person trip planning, proactive protection against flight changes, and in-trip dining recommendations rolling out successively.

On the development side, Fliggy disclosed that during a three-month core development cycle, 148 complex requirements were cumulatively launched, with AI-generated code accounting for over 85%.

From the current product form, Fliggy's focus on travel AI has shifted from answering user questions to integrating with transaction systems. This is a more realistic threshold for OTA platforms entering the AI stage: while model capabilities can iterate quickly, the ability to access real supply, handle complex orders, and take responsibility for subsequent fulfillment still depends on the platform's existing supply chain, system capabilities, and service systems.

Therefore, Fliggy Bangbang is more akin to a transformation of the original OTA workflow rather than the creation of a separate AI entry point. As for how far the "Omni-intelligence Travel Agent" can go, it remains to be seen how stable these capabilities are in the face of large-scale real orders and exceptional scenarios.