Shining a light on customer experience with AI

In a world where customers can research products and compare prices online, buy from any company in the world and influence other shoppers via social media – which Forrester Research has termed the “age of the customer” – businesses must compete on the quality of the service they provide.

“Truly understanding what the consumer is saying and rapidly responding is becoming more and more critical to business survival – artificial intelligence (AI) is key to success in this,” says Catherine Havasi, chief executive of Luminoso.

Not becoming customer-centric has real consequences. Research conducted this year by Accenture reveals that 52 per cent of consumers have switched to another company in the past year due to poor customer service. Once these customers have left, they’re usually gone forever; 68 per cent of customers who have switched say they will not return.

The customer experience industry is changing to reflect this new reality. While companies could previously get away with sporadically collecting and analysing consumer data, real-time analysis is now a necessity. Companies are increasingly turning to artificial intelligence and natural language processing tools to track and address feedback, making it easier for instantaneous understanding and reaction.

Luminoso, a text analysis company, uses AI in conjunction with natural language processing and machine learning to help its clients understand not just the words customers are using, but also their underlying sentiment and intent. The software works, simply, by turning language into mathematical vectors using ConceptNet, an open-source semantic knowledge base created at MIT Media Labs. ConceptNet is part of the AI backbone that teaches computers to “think” like humans.

This semantic application of AI encases the words and phrases that people use, and the common-sense relationships between them. This enables computers to understand accurately and automatically human language and, critically, to learn new words like slang, acronyms and industry-specific terms from the context in which those words appear.

By incorporating AI, the time needed to analyse customer data with a high degree of accuracy has shortened considerably, especially compared with legacy systems that rely on keyword searches and Boolean queries.

In one instance, Luminoso worked with an industry-leading design software company to analyse its live chat data. Luminoso’s software surfaced a small, but growing, subset of complaints. These complaints took many forms, including “empty cart”, “can’t check out” and “products disappear”.

Due to the AI capabilities of its software, Luminoso identified these complaints as symptoms of the same root issue, namely a technical bug with the company’s online payment system that was causing customers’ online shopping carts to empty when they tried to pay.

Luminoso was able to drill down a step further and determine that this problem was isolated to European customers. While the company worked on a comprehensive solution, its customer success team acted quickly to notify European customers about how to navigate the issue. This swift action prevented $11.5 million (£8.7 million) in lost sales.

Another Luminoso client, a leading office supplies retailer, provided rewards in exchange for used supplies. Luminoso’s software began picking up numerous complaints regarding this reward programme. Chat logs revealed many customers were confused about the details of the incentive and were frustrated after trying to exchange supplies that did not qualify for a reward. The retailer was able to take steps immediately to clarify directions on how the programme worked.

Quick responses like this matter. As these clients found, rapidly analysing and taking action on customer feedback can both improve customer satisfaction and prevent lost revenue. Gartner reports that while 95 per cent of companies surveyed collect customer feedback, only a tiny 10 per cent actually implement improvements based on insights from this data.

As Ms Havasi concludes: “If companies are to make the shift towards becoming truly customer centric, they need to work on not only capturing, but also understanding and responding to customer feedback.”

For more information please visit