Domain: Travel and Hospitality
Service Area: AI & ML
Context
During the online bookings of flights/hotels, it is very important to distinguish between legitimate and fraudulent behaviours across the booking funnel.
Solution
The solution implemented inventfor had the objective of detecting any fraudulent booking into the system upfront and block the transaction immediately before it happens.
Unsupervised models have been utilized to spot anomalous behaviour into the system, this model comes in the category of Self Learning AI that continually retrains to detect outlier transactions and behaviour.
While designing this solution, the volume of transactions was also kept in consideration as that’s the key of any such business and system needs to be designed in a manner to detect any outlier in a timely manner to maintain the user experience on the Booking portal.
Outcomes
The solution implemented had achieved in a short time the 100% fraudulent booking detection accuracy before it happens.