The risks of identity fraud remain as high as ever for consumers and banks alike. Global research from Experian shows that 58% of consumers have been a victim of online fraud or know someone who has been a victim, or both.
By Brendan Van Staaden, managing executive of MoData Interactive
The challenge that banks face is protecting their customers from increasingly sophisticated identity thieves and scammers without adding undue friction to customer experience (CX).
One of the most powerful tools institutions can leverage to simultaneously improve CX while outsmarting bad actors is conversational artificial intelligence (C-AI). Today’s conversational AI solutions enable a bank to serve customers at scale and within seconds across the consumer’s channel of choice, whether that’s a voice call, instant messaging, smart interactive virtual assistant response, or a web-based conversational channel.
This technology enables a bank to rapidly escalate issues for large customer volumes without long hold times. It also clears the way for human contact centre agents to focus on the more complex customer questions or sales opportunities, instead of on handling routine calls about credit card limits or bank account charges.
Conversations with twists and turns
Behind the scenes, the AI can interact with thousands of users in a personalised way, with an AI-powered decision-making layer guiding engagements. Today’s technology enables banks to design non-linear conversational scenarios with lots of twists and turns. It’s not necessary to pre-plan all the possibilities of a conversation because the deep learning platform automates all of it.
This allows financial institutions to drive faster interactions, more efficient and accurate outcomes, more personalised service and not least of all, better risk awareness and attention management. Indeed, conversational AI is emerging as one of the more powerful tools that banks can leverage to bolster their protection options against fraud and money laundering.
Managing the risks of identity theft, fraud and money laundering is a major pain point, given the sheer volume of attacks and scams. Authenticating customers and transactions soaks up an inordinate amount of time for frontline employees. The process can be subjective and prone to human error. The regulatory consequences of getting it wrong can be severe and customer experience may suffer when customers are kept on hold for a bank employee to verify their identity.
With conversational AI, the process can be aided by the machine-learning powered system trained on immense reserves of data about customers’ and fraudsters’ behaviour. Such a system is agnostic to the emotional appeals fraudsters might make to get bank employees to grant them access to funds or personal data.
Firm but empathic
Conversational AI can be trained to be empathic and supportive, whilst nonetheless creating an inherently challenging barrier for fraudsters. When it flags a conversation as suspicious, it can refer it to a human agent for help. Furthermore, the AI system can make pointed and relevant decisions within seconds, saving time and frustration for the customer.
The beauty of such a platform is that it will get consistently better at flagging suspicious behaviour while streamlining the processing of legitimate requests from customers. Predictive modelling, combined with machine learning, will deliver faster propensity models, enabling a bank to save time and act more efficiently. Many processes that once needed human intervention can thus be automated.
Machine learning can be combined with voice biometrics and speech recognition to create an even more powerful defence against fraud. The UK’s NatWest uses speech recognition solution screen incoming calls and compares voice characteristics to a library of voices associated with fraud against the bank. The software flags suspicious interactions.
Still early days
We are still in the early phases of using conversational AI to improve the anti-money laundering and fraud management systems in banks. But it’s clear the potential of combining conversational AI with anti-fraud machine learning is enormous. IDC forecasts that banks worldwide will spend $31 billion on AI embedded in existing systems by 2025. Those that fall behind the curve risk failing to meet customers’ and regulators’ expectations.