by Nigel Cannings
The number of fines issued by the Financial Conduct Authority (FCA) more than doubled in 2022. From the breach of Anti-Money Laundering (AML) rules, and Know Your Customer (KYC) checks, to problems with operating guidelines, individuals and businesses within the UK paid out more than £215 million in fines last year. Regulations have become more stringent. Compliance is much more closely monitored, leaving businesses struggling not only to adhere to the regulations but to demonstrate that they have done so and one of the most effective solutions has taken the form of voice tech.
How can voice tech support fintech compliance?
Voice tech has evolved significantly in the last five to ten years. Historically, voice communications were recorded, and a small, random selection of customer interactions (say 3% and 5% of calls) would be reviewed by a dedicated team of compliance officers to listen back and highlight any potential problems. It worked as far as it went. But it left a giant – 97%-of-interactions-sized – gap in which compliance failures could fester.
As speech recognition technology advanced, keyword flagging tech using phonetic search helped enhance the process, allowing businesses to highlight any words or phrases that may potentially be problematic, for faster and simpler investigation but the value of the system was still limited by volume, and its ability to analyse vast amounts of information. New technology has changed that.
The capabilities of contemporary voice tech
- Artificial intelligence (AI). There have been a number of developments in speech recognition technology. Convolutional Neural Networks and other Deep Learning approaches have brought technologies that enable machines to ‘understand’ human speech. Natural Language Processing (NLP) not only accurately transcribes human speech – regardless of accent or language – but allows for the detection of emotion and context. So, if something potentially inappropriate is said – perhaps directing a customer to an unnecessarily expensive product or service – the machine can discern whether it may have been said in jest rather than in earnest.
- Machine learning (ML). ML works in collaboration with AI. Meaning that the initial training data used to program AI systems can be continuously built upon, allowing for increasingly sophisticated language detection and ‘understanding’, improving system performance through each customer contact and supporting the improved identification of vulnerable customers, as well as enhanced fraud detection.
- Enhanced accuracy. Because modern neural networks can filter out extraneous noise in audio signals, businesses have the power to enhance the accuracy of transcriptions, removing the risk of false positives. With the latest Voice Attribution technology, Voice Biometric clustering, individuals can be automatically assigned with a unique “voiceprint”, helping to reduce the risk of fraud, while protecting businesses against misidentification,
- Volume. Contemporary digital technology and storage capabilities mean that businesses are now able to record 100% of all customer communications, and to monitor them in real or near-real time. This means that any problems can be almost immediately identified and dealt with to prevent further perpetuation and limit compliance failures.
- Searchability. Increased data in compliance is only of benefit if you have the means to manage it. With accurate transcription and AI, it becomes possible to search almost limitless volumes of data. Across multiple channels. And often in real-time, making finding evidence for FCA investigations simpler, faster, and more accurate. Call Summarisation can further contract investigations through the creation of easily searchable bite-sized summaries.
- Explainability. While AI is becoming an increasingly useful tool, it is also coming under scrutiny for bias. But with the right AI system, you can build explainability into your processes. This not only allows you to answer any questions when issues arise and work out why specific decisions were made. But it prepares your business for compliance. In the UK and EU, the General Data Protection Regulation (GDPR) states that businesses must be able to explain how automated decisions were made, and this is something likely to be put in place in other countries in due course.
Voice recording has been a mandated part of many electronic communications in the financial industry for years but its remit is growing. It is becoming more capable and useful, not just supporting compliance, but helping to protect vulnerable customers by identifying those who may need more support. It also helps fight fraud and its role now is primarily risk mitigation. The fact it streamlines processes, enhances training, supports growth, and builds customer service, is a beneficial bonus.
The author, Nigel Cannings, is the CTO at Intelligent Voice. He has over 25 years’ experience in both Law and Technology, is the founder of Intelligent Voice Ltd and a pioneer in all things voice. Nigel is also a regular speaker at industry events such as NVIDIA GTC and holds multiple patents in Speech, NLP and Confidential Computing technologies. He is also an Industrial Fellow at the University of East London.