Utilising AI to identify vulnerable customers.

The Financial Conduct Authority (FCA) published a strict set of guidelines highlighting the steps companies should take to understand the requirements of vulnerable consumers and ensure they are treated properly.

Protecting vulnerable consumers is a key focus for the FCA and is more important than ever due to the ongoing impact of coronavirus (Covid-19) and the current economic situation. The FCA guidance aims to help firms treat vulnerable customers fairly and includes examples of how to put this guidance into practice.

The Financial Conduct Authority (FCA) has identified four areas of risk that locate an individual on a spectrum of vulnerability risk. These are health, life events (such as bereavement or unemployment), financial resilience, and capability (meaning financial literacy).

Complying with the FCA Principles for Businesses

The FCA principles require firms to treat customers fairly and make clear what the standards set by the principles mean for firms so that firms understand what is expected of them. It sets out what firms should do to meet those standards. While firms are not bound to adopt or follow any of the specific actions described in this guidance, they must meet the standards set by the principles and treat customers fairly.

Because anyone can find themselves in vulnerable circumstances at any time, the guidance is relevant to firms serving retail customers, including some business customers, regardless of the firm’s size or sector.

Read the complete guide here – FCA

To achieve good outcomes for vulnerable customers, firms should take action to:

• Understand the needs of their target market/customer base
• Make sure staff have the right skills and capability to recognise and respond to the needs of vulnerable customers
• Respond to customer needs throughout product design, flexible customer service provision, and communications
• Monitor and assess whether they are meeting and responding to the needs of customers with characteristics of vulnerability, and make improvements where this is not happening

The Solution

The Saascoms’ collections’ AI engine, ‘Omnireach’, has monitored and analysed millions of conversations and agent responses in the Collections & Arrears environment over the last four years. Designed to assist collections businesses in complying with the Financial Conduct Authority Guidelines, our processing engine has a 93.7% intent and sentiment match success rate.

Features of Omnireach for vulnerability include;

• Natural Language Processing: Intent and sentiment matching.
• layered intent and sentiment, i.e., the bot understands the difference between “I cannot pay you today” vs “I cannot afford to pay my rent”. It will assign intent and sentiment and ensure the conversation is treated as per our client’s workflow and parameters.
• Auto-response based on intent and sentiment, i.e., financial difficulties.

Omnireach can:

• Transfer to a priority agent queue.
• Send an API back to the client’s CRM, placing the contact strategy on hold pending agent interaction.
• Send links to external agencies for help, i.e., Citizens Advice, Payplan etc.
• Depending on intent matching, a digital income and expenditure form can be sent, linked to open banking, providing document upload capability for supporting information. The I & E journey has two-way touchpoints to an agent should the end customer need assistance.

The same approach is also used when identifying other vulnerabilities, i.e., mental or physical health. The bot immediately triggers an API to cease contact strategy and links to external organisations, i.e., The Samaritans. The conversation is then flagged in a high-priority escalation queue.

Language translation – allows customers to communicate in their first language. We identify and translate 92 different languages, enabling customers to communicate in their preferred language and the agents to respond by translating in real time.

Agents can help train the Collections and Arrears AI engine. By copying content from conversations and adding them directly into the AI engine, the Saascoms platform will evolve and ensure AI can identify future conversational meanings.

When a customer is in financial difficulty or a vulnerable situation, three main challenges exist:

• Getting customers to make that initial contact
• Getting customers to make that contact as early as possible
• Providing customers with the correct contact route

Omnireach has an AI function that can not only detect vulnerability but can also match the category of vulnerability, i.e., suicide, self-harm, emotional, terminal illness, substance addiction, bereaved customers, third parties, or financial distress (can’t pay you vs cannot pay rent/mortgage).

By detecting the vulnerability type, the Chatbot makes decisions on how to best handle the customer. For example, someone identified as suicidal or self-harm is immediately transferred to a live agent with specific training for handling this type of customer. The bot also sends an API back to the CRM to stop the contact strategy workflow, offering a chatbot response with independent contact advice, i.e., Samaritans, CALM, Papyrus etc.

Statistics from Q1 2022 suggest 72.4% of conversations engaging with the Saascoms Collection & Arrears Chatbot are resolved on the first contact.