Contextual Intelligence – A new tool in the war against fraud

Billions of friction-filled moments, which risk upsetting a customer’s journey, occur daily, as financial organizations of all kinds—from traditional banks to fintechs and telecommunications providers—contend with making millisecond-quick authentication decisions. Mobile devices aim to make transactions fast and easy, but the mix of underlying data has become incredibly complex. That’s the challenge contextual intelligence promises to help cut through.

Contextual intelligence provides the answer to a bank’s question, “What answer, to what question, do I need to know, at what moment in time in the customer journey, to make sure the transaction is legitimate?”

PSD2 regulation in the UK and European Union also impacts, as it frequently compels the use of strong customer authentication. Online payments, transfers from bank accounts and even changing account details can all trigger an authentication requirement – even when the payment service provider deems it unnecessary. Constant demand for multi-factor authentication is frustrating for customers so when a provider can limit them it’s a win for customer experience.

How do financial institutions benefit from contextual intelligence?

There are three main advantages to working with contextual intelligence.

  • The customers’ experiences will simply be better. Less friction, fewer false positives, higher detection rates of actual frauds and scams and better experiences translate into greater customer loyalty and ultimately, more profits.
  • Cost. By right sizing the quantity of data that is relevant to answering each question, companies can buy significantly less data from external sources, saving costs.
  • Streamlined operations. Delivering more efficient, integrated foundations, contextual intelligence provides a decision framework that can be applied to many customer journeys: account opening and KYC, retail banking money transfers, person-to-person (P2P) transfers, credit applications — anything that requires the orchestration of data from multiple sources to answer point-in-time questions.

Real-world example

A customer goes on holiday in Thailand, she wants to access her online banking to pay a bill but didn’t bring her laptop on the trip, and her phone has run out of battery. She goes to the computer in the hotel lobby to log into her online banking. Contextual intelligence can drive the following sequence, to simultaneously ensure security and a streamlined customer experience.

  • This new device appears risky. The customer is asked to authenticate via an email link since, at the moment, she can’t accept a text on her phone.
  • The customer logs in, and it’s immediately apparent she’s a long way from home, on a computer she’s never used before. It’s appropriate to step up her authentication; she is presented with challenge questions that are successfully answered, and she pays the bill.

Here’s a twist on that scenario: Half an hour before logging on, she made a withdrawal at the ATM in the hotel lobby using her debit card. The stepped-up authentication above would have been bypassed because, although the customer is logging in from a new device and is in Thailand, her bank knows:

  • She withdrew money at an ATM at the same location.
  • That appeared suspicious, so an agent from the bank’s fraud desk called the customer to verify that she was indeed making the withdrawal. The customer verified the transaction, and then her phone ran out of battery.

When the customer logged on to online banking from the hotel lobby, contextual intelligence would have dictated that, since the customer has verified she is in the same geolocation and had used her card, no further authentication would be required to pay a bill. Adding friction to this low-risk activity would be unpleasant for the customer — but if she decides to do a high-risk online banking activity, such as send a large payment from her bank account to a person or business she’s never paid before, the system will ask to authenticate further.

Different transactions require different questions and different answers — and different contextual intelligence.

Doug Clare, Vice President of Product Management, Fraud Protection and Compliance, FICO