Why Banks Need Context and Orchestration to Fight Against Fraud

Fraud can stamp a customer’s experience negatively, but how a bank handles fraud can also have both positive and negative results. In fraud management, failure happens when legitimate customer activity is stopped, customers are uninformed, processes relating to fraud take too long, and customers are treated differently depending on the product they happen to use.

Success occurs when genuine transactions go through, customers can be contacted in real time, customer interactions are both convenient and personalized, and customers are informed and feel empowered. Critically, the margin between success and failure can be very slim.

Increasingly, a lack of context in fraud management will result either in false positives that stop customer activity or limit the ability to detect and respond to emerging scams. Context is important, but it’s the type of context that makes the difference. It’s very easy to focus on suspicious negative context, however positive context is almost more necessary for effective and hyper-personalized fraud management in today’s environment.

Too Much Control Can Be Bad

Some may ask whether there is potential harm in preventing “too much fraud,” and the answer is yes. If controls are too tight and there is too much friction, legitimate activity will be prevented at a greater rate. When this happens, banks risk alienating more customers. In most countries (see chart), significant percentages of consumers would consider changing banks or card providers after experiencing 1 to 3 falsely declined transactions.

Source: FICO Consumer Fraud Survey of 12,000 people, December 2021

And as stated above, when a fraud management solution prevents legitimate customer activity, it should be considered failure. Banks need to focus on decisioning off a broad set of both negative and positive customer context required, from a data analysis perspective, to balance the positive with the negative.

Know Good Context When You See It

Rating context as either positive or negative may seem arbitrary, but it isn’t. There is a gap between what data a bank needs to incorporate into any solution and what questions it must answer, rather than settling for what may have become bad over time and is now insufficient to meet current challenges.

Often, fraud teams are focused on what has become “bad” customer context. This includes addressing simple questions that don’t offer much dimensionality when it comes to analysis. For example, the process may only ask whether the customer has used a product before, when an account was last accessed, whether more than one person uses a given email, whether a mobile number has changed, how many customers use a given device, and whether fraud detection tools like biometric facial recognition or voice print have passed or failed.

All this data is important, but it becomes much more valuable and useful with “good” context added to it. Having the full picture might mean a different treatment or different fraud controls should be used. It’s important to know what:

Add Fraud Segmentation to Drive Treatment

With the addition of “good” context data, banks can not only better address fraud like account takeover and scams, but also respond in hyper-personalized ways to address fraud and other customer experience scenarios. Ideally, any treatment should turn a negative event into a positive customer experience. To do this most effectively, banks need to consider segmenting customers from both banking and fraud perspectives (see chart).

A typical banking perspective looks at customer by net promoter score, status level, net worth, and number of products. A fraud perspective adds important fraud history information to guide hyper-personalized responses to events. That way, a bank has the option to approach repeat victims of authorized payment scams very differently from how they might treat first time victims of a credit card account takeover, for example.   

Orchestration to Automate “Sensible Friction”

Delivering sophisticated responses will depend on a bank’s ability to interlink different treatments and responses, deploy different risk decision and treatment paths, and orchestrate the outcomes of one decision with another.

Orchestration is how banks can move from procedural processes that aim to treat each fraud type differently to a strategy that includes coordinated capabilities and responses, informed by AI and machine learning, to drive better experiences and outcomes for customers while helping stop fraudsters.

When orchestrated treatments are driven by hyper-personalized customer views enhanced with good contextual data, banks have a basis for automating fraud management controls. Orchestration gives them the flexibility to adapt to the changing fraud environment, while providing the smart insights to discern fraud from legitimate customer activity.

This is what we at FICO often call sensible friction and it is a critical part of achieving the proper balance between customer demands for immediate experiences and the need to protect them from fraud within those experience – especially when they have become unwitting victims.

How FICO Helps Orchestrate Fraud Protection and Customer Experience

• Explore FICO’s fraud detection and prevention solutions
• Check out my recent FICO World presentation on how to use hyper-personalization to balance fraud prevention and customer experience
• Learn how you can author and orchestrate impactful decisions with FICO Platform
• Dive into FICO Platform Composition and Orchestration capabilities