Case Study: Global Manufacturer Reduces Contact Centre Costs


A global industrial manufacturing company has one of the largest and most complex contact centres in the world. It receives thousands of interactions every day from engineers using their products in factories.

The Challenge

Most of the interactions are via online resources, although thousands still phone in. Time is of the essence, particularly if a manufacturing line is stopped.

This, combined with the huge range of complex products, means that the company needs a large number of skilled operators in its contact centre and escalation to specialists for third-line support.

The company was looking for technology to assist the contact centre to reduce the time taken for operators to triage the query and retrieve the right information to resolve it.

It was also interested in technology which could obviate traffic to the contact centre by firstly improving online resources, and secondly looking for insight to enhance the usability and, in some cases, the reliability of its products.

The Decision

The company selected PrediCX to do this. PrediCX acts as a supercharged labelling engine which uses a proprietary form of machine learning called Optimized Learning to ‘label’ (aka classify or tag) any customer interaction to a high degree of accuracy. It doesn’t need data scientists to set it up or run it.

It is able to accurately classify incoming queries and match them with the ‘next best response’, be it a resolution or a request for further clarification of the query. A human operator would validate this before replying to the customer.

Accuracy scores for the labelling are also provided so that thresholds can be put in place to automate (when confidence is very high) or provide guidance for humans to choose the best response if there are competing options.

It is easy to plug into any system via an API. As well as dramatically cutting down the time taken for operators to assist, the labels themselves proved invaluable for managers to understand which queries were being asked most often.

This insight was used to enhance and develop their knowledge base accordingly, and also to optimize the online resources, enabling customers to search for the issues in free text, retrieving the appropriate response without the need to contact the contact centre.

In addition, product managers could gain valuable product knowledge from the labels. They could identify what enhancements were needed for usability and reliability, and even predict future issues and prevent them.

One of the critical features of PrediCX is the ability for ‘early warning’, i.e. the ability to autonomously spot new and changing signals. So if there are quality issues, new product launches, or changing customer communications, these are picked up as they occur and the system adapts accordingly.

The Result

The overall result has been a dramatic reduction in time and cost of operations, as well as a significant increase in customer satisfaction, which has had a far-wider positive impact.

The staff at the contact centre are more satisfied, as their time is devoted to more problem solving and the complex, one-off events that machine learning cannot help with.

Happy with the performance and reliability of PrediCX, the company is considering launching chatbots to enhance the benefits further.

Author: Robyn Coppell

Published On: 1st Aug 2019
Read more about - Contact Centre News,

Follow Us on LinkedIn

Recommended Articles

18 Workforce Management Case Studies
Breakfast on a bed in a hotel room
Case Study: HotelBeds Tackles the Challenges of a Global Pandemic
jargon definition
Contact Centre Jargon and Terminologies
Case Study: Contact Centre Improves Customer Satisfaction With AI