Financial IT: Please introduce to us your company and products/solutions you are offering?  

Tom Rimmer: Speechmatics is a fast-growth Cambridge scale-up leading the market with our award-winning any-context speech recognition engine. We have offices across Europe, the US and Asia. 

Automatic speech recognition (ASR) uses machine learning (ML) to automatically transcribe what people say. Our speech recognition technology enables companies across industries to detect and transcribe voice in any context; in real-time or for pre-recorded audio, recognising acoustics, languages, dialects, multiple speakers, punctuation, capitalisation, context and implicit meanings. 

For example, our unique Global English language pack supports all major English accents for use in speech-to-text transcription. Trained on thousands of hours of spoken data from over 40 countries and tens of billions of words drawn from global sources, customers no longer have to worry about being misunderstood. Transcription output can be formatted in American English or British English allowing the transcript to be applied for any use case, whether that is contact centres, media monitoring or bolstering risk and compliance functions. 

Financial IT: What is unique about Speechmatics and how does it stand out among its competitors? 

Tom Rimmer: Firstly, we have mastered 31 languages in a multitude of dialects and accents, regardless of context. Our advancements in accuracy and languages means our technology continues to be adopted by some of the largest blue-chip companies in the world, such as Deloitte UK, ING Bank and Adobe, and widely recognised by analysts.  

No matter the use case, our customers use one platform, whether it’s a bank looking to optimise risk and compliance or a retailer wanting to boost customer experience. They are able to harvest data and understand at scale. This makes the technology versatile and adaptable. 

Control of data is very important for our partners and customers and so our platform complies with data sovereignty restrictions across multiple geographies. It can be used on-premises to ensure data remains within a business’ private environment, with their choice of cloud provider, or on Speechmatics’ cloud offering. Users also have the option to transcribe pre-recorded files or transcribe conversations in real-time. 

Financial IT: The current Coronavirus crisis has posed huge challenges to the fintech sector. In your opinion, how has COVID-19 impacted the RegTech and financial industries, and what are the main changes that you expect to take place in RegTech going forward? 

Tom Rimmer: COVID-19 has pushed the industry into uncharted waters, with financial service institutions having to face new and unforeseen challenges. For example, mass remote working has meant banks have had to adapt to new practises in a short amount of time. Not to mention the increasing risk of staff working in silos, falling ill or being made redundant/furloughed has put a strain on risk management. Less people to man the fort and more inbound enquiries can leave blind spots, especially if they are working in their own bubbles at home. This poses complex regulatory challenges, like the onboarding and managing of customers whilst complying with regulations such as KYC, AML, GDPR etc. In recent weeks we have seen that the SEC issued a risk alert around the concerns of out of office working and possible exposures in their monitoring and supervision obligations.  

During this period of uncertainty and economic downturn, financial institutions are looking to automate their risk management processes. This is where RegTech will really shine. As more customers will be interacting with the banks over the phone rather than going into store, a particularly important component of RegTech going forward will be voice technology.  Through automatic speech recognition, the technology will be able to take on the administrative burden of transcribing the call data into text format, from an already shrunken workforce. This gives staff the ability to efficiently process customer interactions and have visibility of any compliance issues meaning regardless of whether staff are at home or in the office, good governance and due diligence will not be compromised. 

Financial IT: To gain a better understanding of the customer interactions banks need to focus on data. In this context, how can banks enable efficient data extraction and management? 

Tom Rimmer: When it comes to understanding customers better, banks have to be able to analyse all the data they have on a customer. This means voice data too. Even before COVID-19, phone calls remain the most popular choice for communication and in the early stages lockdown saw some banks experience a 400% increase in calls from customers seeking financial support. However, despite a wealth of information being captured on these calls, banks fail to make the most of the unstructured voice data due to resource constraints.  

But with the right automation technology in place, banks can have a better view and understanding of their customers. Voice technology can support banks to visualise 100% of customer calls within the contact centres to streamline processes, flag compliance issues and optimise customer service. Using data to steer conversations can boost first contact resolution, and reduce both customer and agent churn which is very important in the current market conditions. 

To aid in streamlining this, automatically transcribing calls helps financial institutions evaluate and then categorise every customer interaction. These can then be grouped according to specific compliance regulations which can then be addressed appropriately. By recording and analysing every customer interaction, companies can get a 360-view of the customer, map the customer journey and comply with regulations more effectively.  

Financial IT: Voice technologies are developing rapidly in recent years. How can voice technology support banks’ risk and compliance needs?   

Tom Rimmer: Financial services have a lot of red-tape and regulation to contend with. FCA COBS 11.8, MiFID II and GDPR all require banks to have processes in place to better manage, store and process data. For example, under FCA COBS 11.8 banks need to have an agreed method to record interactions across their operations.  

Regulators are also constantly challenging banks to have a comprehensive suite of tools to deliver robust compliance practices. That’s what makes ASR so important: it reduces the overall cost of delivering compliance which is essential to the protection of customers and businesses against fines. Not to mention that a key component of RegTech is voice technology, which provides significant efficiencies for regulatory compliance and customer experience for banks. Transcription is required to transform unstructured voice data from calls into text which can be used alongside other text-based information and be triaged by analytics and other purpose-built RegTech solutions.  

From a risk and compliance perspective, unstructured voice data opens up gaps and blind spots if businesses do not have visibility of it. Voice technology is vital to enable monitoring of voice interactions at scale, identifying data breaches and mis-selling by transcribing all audio files for analysis. If businesses can capture and analyse the voice data from their calls or audio files efficiently, investigations can happen quickly and easily which ultimately saves businesses time and money.  

Financial IT: To conclude, what upcoming news do you have at Speechmatics that you can share with our readers? What is next for Speechmatics?

Tom Rimmer: At Speechmatics, we’re continually pushing the boundaries of ASR and taking on the big players in the market in terms of the level of accuracy and languages. Innovation is at the heart of what we do as a company. With 14 PhDs in machine learning and speech recognition we continue to develop our ASR technology, adding more languages, dialects/accents and features. 

Speechmatics is currently working with many partners to address multiple use cases for improving customer journeys/insights, retention and proactive information governance policies, monitoring and surveillance of traders for market abuse etc and investigation/early case assessment solutions.   

With the world still in a state of lockdown, and face-to-face interaction minimal, many industries have experienced an influx of customer calls via call centres. The need to be able to efficiently and effectively process and manage customer conversations is very important to avoid customer churn, any damage to brand reputation and to stay compliant with evolving regulations.  We recently partnered with Prosodica, (subsidiary of Vail Systems) to help them analyse billions of customer interactions at scale, regardless of dialect/accent to deliver better caller experiences. Voice data transcribed into text format means contact centre agents can gain insights to derive meaning from customer interactions to boost customer satisfaction.

Tom Rimmer is a Senior Sales Executive with over 20 years of industry experience. Tom has a diverse background in compliance, eDiscovery and information governance. Tom has notoriously worked with large multinational banks on software-based solutions throughout his career.