19th July 2022

AI in contact centre authentication

By Grant White

For the last two years, lockdown restrictions have had a profound impact on contact centres. With consumers unable to deal with enquiries in person, contact centre agents have been flooded with a tidal wave of new customer enquiries. At the same time, managing this demand against the backdrop of remote working has been a challenge. 

But in response to this challenge, a revolution has been brewing in contact centre technology, with artificial intelligence (AI) sitting right at the centre of this trend. But how is this technology being used – and what are the challenges that it might pose for organisations in the next few years.

A new breed of contact centre technology

Since Covid, contact centres have increasingly focused on improving the customer experience and operational efficiency. In fact, recent research from The Harris Poll and Interactions found that most of the call centre technology buyers they surveyed (53%) had cited improved CX as a primary driver for their AI implementation decisions. At the same time, a report from MIT suggests that AI lets contact centres process higher call volumes in about 80% of cases. 

The case for AI in the call centre is clear. But what technology is being used to power this revolution? These three are some of the most common:

  • Predictive call routing (PCR)

Predictive call routing uses complex AI algorithms to analyse real-time call data, such as a caller’s preferred communication channel, their call history or their personality profile. This helps match the customer and the situation to the best available agent, providing a better experience for the customer and improving contact centre handling times, transfer rates and abandonment rates. 

  • Interactive voice response (IVR)

IVR systems are some of the most ubiquitous technology available in call centres today. They use pre-recorded messages and text-to-speech to interpret the intent of the call, verify the caller’s identity and permissions and carry out basic authentication. This reduces the time and money that is spent when live contact agents have to undertake these basic tasks. 

  • Conversational AI

The latest conversational AI technology uses natural language understanding (NLU) to enable conversations between the system and the customer, which is similar to text chatbots that can be found on many websites. This can provide a faster way to manage calls, rather than relying on contact centre agents, and has the added benefit of offering multi-language support. 

The case for employing the latest AI solutions in call centres is compelling, offering more effective call routing, helping firms to service a range inbound queries at scale, and providing a much more flexible way of dealing with customers’ diverse needs. Though as we’ll explain, AI presents as many challenges for contact centres as it does solutions.

A gap in the fence

The usage of AI in the contact centre has clearly developed in the last few years. But at the same time, so has the technology that fraudsters use to target them. 

Increasingly, fraudsters are finding more and more sophisticated methods of penetrating contact centre defences. Some of these involve using AI themselves – but many simply involve duping the technology and techniques that contact centres themselves have employed. Despite their advances, modern contact centre technology still has its limitations – particularly when it comes to identifying subtle, low-signal fraud activity. 

One particular example of this is voice phishing or vishing, which is the process of gathering information about a customer or fraud target through the telephony system. The fraudster might pose as the customer, and ask to check the customer information on record such as their address, telephone number, social security details or banking details. Fraudsters then use this to successfully commit fraud (often not even through the contact centre) or sell this information to other fraudsters for the same purpose.  

Traditional IVR systems often use a combination of voice inputs and touch-tone keypad selections to assemble the responses they need. Fraudsters regularly exploit these systems to gain and validate stolen information that can help them commit fraud. 

But it’s not just IVR systems. Deepfake audio, mimicry and real-time voice masking are becoming increasingly common ways of accessing accounts, extracting data and, ultimately, stealing money. These will become more and more common as this technology develops. 

How to stay protected, both now and in the future

If you’re a contact centre, there’s every reason to continue adopting new technology to increase operational efficiency and upgrade the customer experience. But it’s important to remember that each new technology comes with new vulnerabilities – and no method is foolproof. 

To stay safe, the best solution is to layer multiple types of fraud prevention technology together, ensuring the best possible level of protection. These solutions could include:

  • IVR authentication
  • Voice biometrics
  • Knowledge-based authentication (KBA)
  • Fraud prevention technology

To discover more about the changing situation of AI in the contact centre, and learn more about these fraud prevention solutions, check out our recent eBook: Authentication with automation.

New call-to-action