Fighting fraud with voice technology

Article contributed by Nigel Cannings, CTO, Intelligent Voice

Fraudsters have become increasingly sophisticated in recent years. From international criminal gangs targeting the life insurance industry to the rise of social media phishing for insurance scamming purposes.

In the light of the cost-of-living crisis, some pundits have begun talking of a pandemic of fraud, with hard-pressed people seeing insurance fraud as the simplest way to survive the current economic constraints.

For insurers, this means increased pressure to keep pace with the scam techniques and technologies and the call centre has historically been the most vulnerable point of contact. However, advances in artificial intelligence (AI) technology combined with speech recognition have opened the door to a new era of audio and video data fraud detection.

Speech AI

One of the leading benefits speech AI brings to fraud detection is consistency. The call centre model is devised to serve as many customers as possible; with customer data stored in a central hub there’s no need for individual case management. This leaves the door wide open to fraud.

With voice technology, that is no longer the case. Capable of detecting and monitoring a wide range of speech features – tone of voice, emotion, speech patterns, language – the technology can be used to uncover any inconsistencies presented by claimants, which may indicate fraudulent behaviour. Working in real or near-real time, it can do this without disrupting service or damaging genuine customer relationships.


Natural Language Processing (NLP) is one of the most important recent developments in voice technology. Devised to help computers make sense of the language humans use day to day, NLP enables the detection of fraudulent behaviour by tracking inconsistencies and irregularities.

However, until relatively recently, NLP systems have been based solely on text. But with the introduction of highly sophisticated speech recognition technologies, it has become possible for NLP algorithms to work with spoken interactions, meaning that speech analysis can be applied within a call centre setting.

The LexiQal behavioural analytics platform, which was developed alongside law enforcement experts, was developed to uncover fraudulent behaviour from the first contact. Running throughout every customer contact, the technology highlights and records story changes, speech inconsistencies, unusual behaviour and any curious language in the background.

This creates a permanent record for fraud analysts to listen to later.


But for all the fraud prevention benefits that AI and voice technology bring, it still has more to give. The speech pattern and key word detection that is valuable in identifying fraud can be used to more positive effect too, through the identification of vulnerable customers.

By recognising that a caller is confused, uncomfortable, or distressed, the technology allows the business to tailor its service offering, perhaps passing the caller on to a more experienced supervisor or taking further steps to ensure that the customer truly understands before processing their claims or purchases.

Additionally, there is real scope for the deployment of the tech to support regulatory compliance and client protection. Providing a clear audit trail that can be easily searched, the technology not only simplifies the keeping of in-depth records but facilitates the easy collection and submission of evidence, should an investigation ever be required, helping to protect both the business and the customer.


The role of anti-fraud technology is simply to protect. It protects the business, employees, and customers against the impact of scammers. While there is a belief that fraud can be a victimless crime, it costs billions per year.

In 2021, insurers detected 89,000 dishonest insurance claims valued at £1.1bn in the UK alone and it’s estimated that almost as many went undetected. The impact on the public can be devastating.

One in 20 motorists under 35 have apparently admitted to having ‘staged’ or invented an accident and claimed successfully. When you move onto a professional level, where scam payouts lead to money laundering and other areas of crime, the societal impact is beyond scale.

Speech AI can help to prevent this. It is easy to implement, cost-effective, and highly efficient, helping insurers to save money, keep quality records, adhere to compliance standards, and protect their customers. And this is just the beginning of its journey.

This article was provided by Nigel Cannings, CTO, Intelligent Voice, a global leader in the development of proactive compliance and eDiscovery technology solutions for voice, video and other media.


Upcoming Event

Motor Claims News

ILC Breakfast News – Wednesday, 22 May 2024 

ARC360 | Claims Essesntials | Claims Tech | Home | ILC News | MGA | Motor | Networking Lunch | New Generation


ILC Industry leader interview: Mark Burton, Partner, Kennedys 

Claims Essesntials | ILC News | Motor


Solera and Audatex agree sustainability partnership 

ARC360 | ILC News | Motor


Insurer Partners


Industry Body Partners

Insights Partners

Motor Corporate Partners

BHR Assist Insurance Services