By Tom Rimmer, Senior Executive at Speechmatics and George Tziahanas, Managing Partner at Sohosquared

Financial markets are under considerable pressure at the moment. Even with lockdown restrictions easing, banks are still urging customers to use online apps and avoid branches. Although the trend of moving towards online banking is nothing new with 49% of consumers preferring it, Covid-19 has been a catalyst pushing that number up to 57%.

There is the increased burden of data visibility across a host of new channels, while ensuring compliance is met to contend with multiple regulations. In 2019 alone, The Financial Conduct Authority (FCA) issued almost £400 million worth of fines in the UK for compliance, legal and governance related issues. With more regulations to contend with, banks and financial institutions could lose more than just being slapped with a penalty/fine for non-compliance. The trust of the customer is also at stake. Two in five customers leave banks after a bad experience, and 45% of those will actively discourage others from using that bank, according to PWC.

Additionally, with more people managing their finances online, communication technologies need to ensure the level of customer service is not impacted by this transition. Even in the digital age customers still want to speak to someone when managing their finances and phone calls remain the most popular channel of communication – especially when something goes wrong.

With the rise of customer calls, the burden to understand, manage and protect customer voice data is more important than ever before. Early into the Covid-19 lockdown, Banking & Payments Federation Ireland (BPFI) said member banks were experiencing a 400% increase in calls seeking financial support, including an average of 7,000 calls a day from customers on mortgage-related concerns.

The burden of manual transcriptions 

Financial institutions produce reams of unstructured voice data but often fail to make the most of it. The challenge is how banks can leverage this information at scale. In their current audio format, there is a lack of visibility of what is discussed and the context of customer and agent calls.

Historically when calls were recorded, compliance teams would manually review the calls for anomalies which is expensive and time consuming. Roughly, every hour of audio recording takes approximately three hours to review. That’s a lot of work and admin for contact centre agents if you consider they experienced a 400% uptick in calls. This makes the task of identifying and flagging non-compliance issues monumental, opening them to blind spots and errors.

Banks also end up analysing only a small percentage of voice data. Research from Call Centre Helper suggests that contact centers across industries analyse less than 3% of interactions. This leaves 97% of call data untouched. The solution lies in automatically transcribing calls. This takes away the heavy lifting for contact centre agents, instead they can focus on indexing and spotting keywords to flag up compliance issues, and identify data breaches and mis-selling.

A good lesson which illustrates this is the PPI scandal. Despite the PPI deadline passing last August, the industry is still suffering from the fallout of PPI claims. The issue stemmed from the fact that the majority of the complaints were based on sales made over the phone or in person, offering little to no record of customer engagement.

The power of automatic analysis of voice data

Using voice technology means that contact centre agents no longer have to manually transcribe calls, tag or categories, but rather can focus on analysing the voice data to help steer the conversation. For example, if certain questions are raised by customers, relevant resources can appear on the agent’s dashboard to optimise and fast-track the resolution.

There also needs to be a single platform which captures all data, irrespective of what channel customers use so contact centre agents can gain a 360 view of their customer. This is not just important for customer experience, but also essential for compliance to have visibility of all of the data held on the customer whether it was over email, chat or phone. This means contact centre agents can have access to every customer interaction over time in one place which can enhance how they respond to customers in the future.

Not only this, but by capturing and analysing the voice data from calls efficiently, investigations into queries can happen quickly and easily, which ultimately saves banks time and money. There is an increasing need to identify any issues which arise during the call This means that feedback forms do not need to be sent to customers after the fact, only adding to the burden of capturing that information.

Whilst real-time call analytics do not completely solve compliance problems instantly, it provides an opportunity to proactively highlight issues immediately either by hints or even by rerouting the call to a supervisor before it escalates. These prompts will in effect improve customer interactions, positively impacting customer satisfaction levels, as issues will be dealt with quickly boosting the score for first call resolutions and reducing the average handling time.

Accurate captioning for regulatory compliance 

Banks and financial institutions work in heavily regulatory environments. They have to be able to evidence they are compliant to the auditor and regulators, grappling with multiple regulations such as COBS 11.8, MiFID II and GDPR. These require them to make all call records transparent and accessible for both regulators or customers. For example, under FCA COBS 11.8 banks need to have an agreed method to record interactions across their operations. By transforming these voice recordings into text through transcription, banks can store and index customer records automatically, making them more accessible.

Research found that 91% of respondents believe increasing investment in contact center compliance software should be considered a priority in 2020. 83% of contact center professionals also said their organisation’s efforts towards customer privacy and private data safety needed to be improved.

Accurately captioning data from phone calls is very difficult – but by being able to overcome this challenge through any-context speech recognition technology, banks can ensure they have more visibility of what is being discussed on those calls at scale to minimise compliance issues and better deliver customer service.