Editorial & Advertiser Disclosure Global Banking And Finance Review is an independent publisher which offers News, information, Analysis, Opinion, Press Releases, Reviews, Research reports covering various economies, industries, products, services and companies. The content available on globalbankingandfinance.com is sourced by a mixture of different methods which is not limited to content produced and supplied by various staff writers, journalists, freelancers, individuals, organizations, companies, PR agencies etc. The information available on this website is purely for educational and informational purposes only. We cannot guarantee the accuracy or applicability of any of the information provided at globalbankingandfinance.com with respect to your individual or personal circumstances. Please seek professional advice from a qualified professional before making any financial decisions. Globalbankingandfinance.com also links to various third party websites and we cannot guarantee the accuracy or applicability of the information provided by third party websites.
Links from various articles on our site to third party websites are a mixture of non-sponsored links and sponsored links. Only a very small fraction of the links which point to external websites are affiliate links. Some of the links which you may click on our website may link to various products and services from our partners who may compensate us if you buy a service or product or fill a form or install an app. This will not incur additional cost to you. For avoidance of any doubts and to make it easier, you may consider any links to external websites as sponsored links. Please note that some of the services or products which we talk about carry a high level of risk and may not be suitable for everyone. These may be complex services or products and we request the readers to consider this purely from an educational standpoint. The information provided on this website is general in nature. Global Banking & Finance Review expressly disclaims any liability without any limitation which may arise directly or indirectly from the use of such information.

How can banks use AI to spot customer vulnerability

By Dan Somers, CEO of Warwick Analytics.

Since the Financial Conduct Authority (FCA) published it’s ‘Guidance for firms on the treatment of vulnerable customers’, regulated firms need to work towards a corporate culture that identifies and supports vulnerable customers. Ideally front-line staff need to be able to spot different types of vulnerability and support them accordingly.

This is no mean feat seeing as it’s believed that up to 50 per cent of the UK population may be exposed to one or even several recognised vulnerability drivers.

There are a number of warning signs that an advisor can pick up on to identify if the caller on the other end of the line is vulnerable but without extensive and proper training these are difficult to spot and the volume of communications is just too much to carry this our manually.

However, the latest in speech transcription and text analytics, supported by AI technology, is now able to automatically detect vulnerability and hints at vulnerability from conversations with customers.

Conversations with customers across all channels, including telephone, email, live chat and social media, are automatically analysed by machine learning models that can detect not only the topics in the queries themselves but also emotionally-driven comments that indicate vulnerability from a basic lack of understanding, likelihood of a disability and circumstances. These vulnerabilities are flagged to the relevant members of staff for action.

The technology works by examining the specific language used by customers against pre-defined phrases and sentiments that may suggest any element of vulnerability. Example concepts or sentiments could be ‘I can’t pay’, ‘I’m having trouble paying’, ‘I can’t read my bill’, ‘I can’t understand the letter you sent me’, ‘I can’t hold on all day’ along with signs of agitation, asking for repetition (a sign that the customer is not retaining information), signs that the consumer has not understood or signs of confusion, mention of medication and so on.

You can then use that insight to trigger alerts for managers or nominated individuals – for example, if a firm has a dedicated vulnerable customer team, reroute specific interactions to key team members, to address customer issues swiftly, make sure you capture every instance of possible vulnerability or integrate with knowledge management systems, so teams can access relevant pre-scripted answers to accelerate the right level of care.

As well as helping to spot vulnerabilities and handle each conversation more effectively, the analytics can also deliver powerful insights for service and process improvements. Firms can personalise training for frontline staff, by focussing on the topics most common or difficult to deal with. They can also identify failures in self-service platforms or other channels to help customers help themselves more. Regulated firms can accurately understand the drivers behind the vulnerabilities so products, services and communications can be reviewed accordingly.

And of course, because every instance of vulnerability is identified and recorded a firm is able to prove to a compliance team that they are taking the steps required to spot and aid vulnerable customers.

One such technology is from customer data specialists Warwick Analytics who are working with several organisations to provide an early warning system for organisations based on the conversations that customers have with the organisation’s staff.

How a regulated company is using AI to spot early warnings of vulnerability

A leading regulated enterprise already had stringent vulnerability and compliance processes. They were using their structured data, particularly transactional data, to ascertain patterns of potential vulnerability.

However, from a more detailed analysis of conversations using AI and machine learning, Warwick Analytics was able to show that the level of potentially vulnerable customers was 5x more than the company initially thought and also included markers of where existing processes may have been at risk of being breached. It was also then able to triage these issues to the most appropriate staff.

In all there were eight classes of vulnerability identified for customers – from their situation, intentions and state. These enabled different classes to be handled in the most appropriate way. This was good for the customers, good for the overall experience and good for the business.

Organisations can set automate responses and trigger alerts that will ensure that not only are vulnerable customers being spotted earlier but are being offered the appropriate level of support.

By implementing the technology organisations can remove huge amounts of pressure from agents and clearly demonstrate their compliance to regulators.