As in many professional fields (medicine, law, construction, etc.), there are generalists and specialists. The same applies to the development of behavioural analytics; there are general modelling techniques and highly specialized techniques. For example, your communications provider may use a behavioural model to spot indications that you are likely to leave their service within the next 90 days. Or an online gaming site may employ general behavioural analytics to identify users in need of intervention.
While human behaviour may appear to be a universal concept, the ability to effectively detect anomalous behavior in one domain is not an indication of effectiveness in another.
In fraud detection, artificial intelligence relies on raw data as well as predictive characteristics that serve as inputs to a model that produces a score. These characteristics represent the inferred patterns or relationships within the data, that are often discovered with machine learning. Data scientists with deep knowledge of the fraud domain then improve this discovery process by evaluating and refining the weights, portions and combinations of predictive characteristics for optimal performance of fraud analytics.
The Importance of Domain Knowledge
Many providers of fraud analytics choose to ignore the importance of domain knowledge in the model development process. Instead, they rely on generic behavior models that must learn to identify patterns of fraud slowly over time, based on relatively few cases.
Consider this example: A 42-year-old woman from KwaZulu Natal needs to bury her 80-year-old mother, using a known high-risk funeral parlor. She is claiming on multiple policies within your organization. Your fraud system has less than a second to make a risk determination.
Is this anomalous behavior? It may be. That is relatively easy to determine. But is it indicative of fraud? That is a tougher question that only specialized fraud analytics, honed on huge quantities of data, can accurately assess.
To maintain a positive client experience, specialized fraud analytics must be used to assess the “tough” questions. This is where advanced profiling, fraud-specific predictive characteristics, and adaptive capabilities separate themselves from generic behaviour analytics.
In a world of real time payment processing and rapidly changing consumer preferences, generic behaviour models are not sufficient for cross-channel, enterprise fraud solutions. After all, when and how someone chooses to manage their insurance profile is not as predictable as his or her likelihood to cancel a fitness club membership.
There are significant financial and reputational risks when appointing a generic behaviour model at a fraud use case.