2004 C.I.P.A. Winners

(back to 2004 C.I.P.A. Winners)

Bell Canada: Emily – Bell Canada’s Voice Recognition for 310-BELL Customer Care

 download pdf version

Advanced natural language voice recognition system answers the call for improving and simplifying automated customer service.

Bell Canada, a national leader for communications in the Internet world, provides connectivity to residential and business customers through wired and wireless voice and data communications, high-speed and wireless Internet access, IP-broadband services, e-business solutions and local and long-distance phone services. Bell Canada is owned by BCE Inc.

"Emily" is an advanced speech recognition natural language application for Bell’s primary customer care line, 310-BELL. By replacing the traditional touch tone IVR (Interactive Voice Response) system with the Emily call steering application, customers can simply say what they are calling about and be routed to the appropriate self-serve application or customer service agent as required.

Challenges

Who among us hasn’t been aggravated by touch tone menus that confuse and confound? Whether old or young, computer-savvy or technically challenged, we’ve all had the frustrating experience of not reaching the right person when calling customer service. In Bell’s case, its traditional touch tone IVR system was in the top 10 customer complaints for 36 consecutive months. Something had to be done!

Objectives

The three primary objectives were to:

1. Simplify the customer experience: The goal was to reduce the number of steps it took a customer to get to their end destination, making it easier for the customer to do business with BELL and access the information or customer service agent required;

2. Reduce misdirected calls: Due to the many specialized customer service representatives across Bell Canada Enterprises, quite often customers would "zero out" from the system and have to be manually directed by an agent to another department;

3. Increase self-serve usage on the IVR: With approximately 200 self-serve applications on the platform, it was very dif. cult for customers to reach the self-serve application they were seeking. As a result, many applications were highly underutilized.

Solution

Bell Canada has dramatically altered the situation for the 7.3 million residential customers directing their inquiries to 310-BELL. Emily is the new voice recognition system on the 310-Bell IVR system in Ontario and Quebec. "She" is the bilingual persona who listens to customers’ requests and then directs their calls to the correct IVR application or customer service centre. At long last, Emily replaces the need for customers to listen and interact with a complex, touch tone, menu-driven navigation system.

Now, when a customer calls 310-BELL and inputs his/her telephone number, Emily answers the call. Emily greets the customer with a friendly prompt of, "Hi, this is Emily, your automated service representative. Simply tell me what you’re calling about, and I’ll direct your call... So, what can I help you with?" Emily is not only delivering a simplified customer experience, she is also helping to improve productivity in Bell’s call centres. With the improved call routing, Emily helped to reduce misdirected calls by 45%, and increase the usage of the self-serve applications by 35%. Cost reductions in 2003 alone were $6.6 million.

In a unique alliance, Bell Systems & Technology partnered with both IBM Canada and Nuance Communications Inc. to develop this application. There were two parts to the successful deployment of this technology. The first was to ensure that Emily understood what customers were saying; and second, to ensure that the right business rules were invoked in order to route the customer calls correctly.

To teach Emily how to listen to customers, approximately 30,000 random customer requests, called "utterances", were captured during the pilot stage of the project. Each utterance is tagged to a list of token destinations representing the desired end state based on the request by the customer. This creates a statistical probability of utterance variations and if it is detected again, the system will place a confidence score on the utterance. Depending on the score, the call will either be routed directly to the end state, or a confirmation prompt will be played to ensure accuracy. As a last resort, the customer is prompted to repeat their inquiry.

Innovative Use of Technology

Only recently has voice recognition technology matured to the point where it is now possible to implement a system that can handle responses to open-ended questions. Customers now have the freedom to use their natural language to describe the purpose of their calls. A customer can say, "I’m moving" or "What is Call Waiting" or "I have a problem with my bill". Emily then reacts to the customer request either by transferring the call to the appropriate self-serve application or contact centre, by prompting the customer for confirmation, or by providing another menu. Rather than the IVR telling customers what they can do, Emily allows customers to speak in their own words and then be quickly routed to the correct destination.

Based on the successful deployment in its internal customer care centre, Bell has now created a product for sale to large business customers. Hosting speech solutions has created a new revenue stream at Bell, which is now consulting and providing assistance to many companies as they turn to speech solutions to improve contact centre efficiency.

A 2004 CIPA Winner!

For its exceptional and innovative application of Information Technology to solve real-world business problems and bring greater benefit to all its stakeholders, Bell Canada was awarded a 2004 Canadian Information Productivity Award of Excellence in the Customer Care category.

(back to top)