The data revolution is all around us. There is no industry that is immune; there is no task or process that cannot be made more efficient by data analytics. Ecommerce fraud prevention is no different. What is changing is the sophistication of the data crunching via Artificial Intelligence (AI) — and increasingly, the perceived role of human intelligence and intuition in the fraud prevention process. But while the promise of AI based data modeling is both necessary and strong, it is not absolute.
You would be forgiven for not realizing this is the case.
Vendors are failing the merchants who turn to it for comprehensive fraud prevention. Content and conversations in the space are awash with solutions discussing how they are using the data crunching capabilities of AI and other forms of technology like bio-metric security to protect merchants. Yet when it comes to human intelligence and training or hiring eCommerce fraud analysts — something that is constantly spoken about as relevant even in an AI powered world — there is little practical advice. This needs to change.
Ecommerce Fraud Prevention Vendors Must Do More to Educate Merchants
In the past, rules based solutions revolutionized how merchants protected themselves. This was the first data revolution; it offered the ability to review common characteristics of fraudulent orders, and either instantly decline them or send them on for manual review when computer models could not make a confident prediction.
By definition, human fraud prevention analysts played a large role in this process. Yes, patterns could be determined by machines but the static nature of rules required human analysts for edge cases; humans were also required to review analytical models when chargebacks spiked. All of this forced solution vendors and industry thought leaders to take the role of the human analyst seriously. Merchants benefited from discussions about the intricacies of order evaluation.
Today, the data revolution — the use of AI — has all but crowded out this kind of discourse. Industry leaders constantly state that human fraud analysts still have a role to play when speaking with the media. But they don’t back this assertion up with any meaningful action. Vendor marketing materials constantly emphasize the importance of technology to the near exclusion of any other factor.
If vendors believe human fraud analysts are necessary for effective fraud prevention, they owe it to the merchants they serve to provide guidance on how to improve their internal teams.
The Human Fraud Analyst Is Dead. Long Live the Human Fraud Analyst.
One reason the human fraud analyst is dead is complexity. Complex AI solutions require vendors to invest large sums in educating merchants about the new threats — increasingly prevalent and well known thanks to dark web fraudster forums — and explaining how new technologies combat them. Solutions providers do this important work well.
Human fraud analyst training requires the same level of complexity. Across the board, industry experts agree that human fraud analysts still matter, making the lack of investment in educating the market even more perplexing. If human fraud analysts remain essential to effective fraud prevention, then educating them on how to best do their jobs is a must. Education in the people will lead to better use of the tools, which will lead to better overall results.
Vendors already possess this knowledge and expertise. In-house fraud prevention analysts work for the best eCommerce fraud prevention solutions to constantly adapt machine learning and AI modeling. Solutions must begin to think beyond competing with one another on technology alone and look for opportunities to educate human fraud analysts on aspects of prevention that cannot be outsourced to a machine.
The human fraud analyst is not dead. Fraudsters continue to evolve ever more complex strategies and tactics. Vendors already on both propositions, and so should agree on a third — the need to not only provide the technological tools, but also the framework and education for the humans behind them.
~The Editorial Board