AI revolutionising CX: Richard Bassett on identifying and supporting vulnerable customers

Traditional approaches to identifying vulnerable customers have often been time-consuming and error-prone, leading to some individuals being overlooked. 

With AI, organisations can automate the identification process and offer targeted coaching to frontline staff, enabling them to deliver exceptional service with empathy and understanding. 

In conversation with MyCustomer, NICE's Richard Bassett explains how AI revolutionises CX, drives personalisation, and advances care for vulnerable customers.

How can organisations effectively identify and care for vulnerable customers using AI?

When AI is integrated across a business, it can analyse every interaction across every channel in real-time. AI can be used to consistently and accurately identify, classify and report on vulnerability risk for every interaction based on the FCA drivers – a key requirement of the new FCA Consumer Duty regulation. It can automate remediation workflows by vulnerability type and track resolution. It can also prevent future mishandling of vulnerabilities by upskilling frontline staff with automated quality and performance metrics and targeted coaching. It drives businesses to achieve FCA compliance, deliver exceptional care in a time of need and empower employees to deliver excellent service. 

AI can be used to consistently and accurately identify, classify and report on vulnerability risk.

What challenges do contact centres face when identifying and handling vulnerable customers, and how can AI help overcome these challenges?

Traditionally, organisations have relied on human agents to manually identify and manage vulnerable customers, an expensive and tedious process. Relying on human agents alone to identify vulnerable customers guarantees that some will be overlooked. Without AI, this is a time-consuming, complex task with lots of room for error. AI not only analyses every customer interaction to identify vulnerable customers, but also simultaneously coaches and guides agents on how to address complex needs in real-time. It eliminates manual and subjective analysis, reduces costs and is highly scalable. It eliminates potential for human error, including biases and weaknesses of human judgment, and greatly increases the level of care businesses can provide to their customers. 

Relying on human agents alone to identify vulnerable customers guarantees that some will be overlooked.

AI can also proactively identify agent skills that are impacting compliance, delivering focused coaching or best practices to prevent future violations. AI can identify broken processes and reduce employee variability to truly transform CX. 

How can AI be used to measure customer vulnerabilities?

AI provides a rich, data-informed analysis of 100% of customer interactions, not only identifying vulnerable customers but providing further details about customer vulnerabilities. AI classifies interactions by FCA drivers of vulnerability, escalations and sentiment to ascertain the level of risk, allowing businesses to prioritise high risk interactions. AI monitors trends by volume, type, location and more.

AI provides a rich, data-informed analysis of 100% of customer interactions.

Proactive analytics can auto-correlate drivers of vulnerability topics and trends across channels, flagging emerging or recurring issues with early-warning systems. This data can be consolidated and reported on a vulnerability dashboard, giving businesses complete visibility of the insights generated by AI. 

Can AI enable agents to manage vulnerable customers better, or does the implementation of AI mean personalisation is lost?

AI augments the level of personalisation businesses can provide to their vulnerable customers. AI identifies signs of vulnerability that human agents otherwise might not pick up. Using this insight, AI can then help to automatically connect vulnerable customers with the most appropriate agent or team that possess the skills and resources to help them - allowing companies to immediately respond to high-risk or sensitive interactions in the right way. AI also goes one step further and can help to upskill frontline staff by providing real-time guidance with how to engage in a manner that cares for the customer and leads to a successful resolution. It greatly improves the management of vulnerable customers. 

What steps should organisations take to ensure that the AI models used to identify vulnerable customer are reliable, unbiased, and ethical?

Organisations should only consider using purpose-built AI that is CX-specific and uses proper brand constitution. AI must be customisable with guardrails, considering specific brand language. Organisations should use an AI system with domain-expertise, trained on billions of historical recorded customer interactions. AI that meets these requirements will be able to generate reliable, unbiased and ethical outputs. Without these specifics, there is the potential for inaccurate or inappropriate responses. 

AI must be customisable with guardrails.

Could you provide some examples of how organisations have successfully implemented AI to improve CX for vulnerable customers?

A motor finance company implemented AI to identify customer interactions with the highest probability of a vulnerable conversation. Through using AI, the company also uncovered a large percentage, 23%, of resilience vulnerabilities, which were previously unknown. AI measured positive correlation between agent behaviours and impact on vulnerable customers. The business was able to provide coaching to its agents to help them become more empathetic and build greater rapport with customers. This business measured an increase in sentiment scores across the board across vulnerable customer interactions.