Every transaction counts, so fraud rules can’t be too tight!

Travel suppliers and intermediaries are attempting to revitalize their respective businesses, offering discounts and coupons to boost consumption. This means every transaction counts and a risk-averse mindset – i. e. rejecting a transaction than to risk passing a fraudulent one – won’t work, wries Ai's Ritesh Gupta 

25th June, 2020


A major lesson from the Covid19 pandemic when comes to balancing UX and security is to make the most of available data.

Do away with a mindset that is commonly associated with rule-based systems, which is built with hard rules or buying limit.

“Every transaction counts and fraud rules can’t be too tight,” Microsoft’s Sondra Feinburg told Ai’s Ritesh Gupta in a recent interview.



Some key points:

  • Data makes difference. As Braintree asserts, more relevant data facilitates more automated learning that can be applied to risk decision-making.
  • Focus on separating a “good customer from a bad customer”. Too much authentication in shopping can upset the shopper. Strike the right balance between payment fraud protection and approval rates.
  • Evaluate a multidisciplinary approach that combines big data, predictive analytics and real-time machine learning.
  • Ongoing monitoring is essential – Machine learning models need to be trained, tested and re-trained as fraud trends evolve, points out ACI.
  • More customers are preferring contactless payments. Work on authentication strategy for a mobile experience.
  • Only dealing with rules have a limited scope, as they are reactive and tend to be about fraud to happen basically. Have a layer of adaptive AI in order to learn the patterns.


Follow Ai on LinkedIn: https://www.linkedin.com/company/airline-information/?viewAsMember=true