Matching Experience with Expectation – Using Speech Technologies to Power up a Digital Workforce

Jun 12, 2017

April and May were big months for the analyst circuit where industry analysts get the deep dives into solution provider’s performance and product roadmaps. There were overlapping events in fact, forcing the analyst community to choose where to go and who to see. But just choosing a handful certainly enabled us all to see which themes were reverberating throughout the industry.  I’m going to pick on just one here as it one of the most interesting developments to date; and that is the emergence of Virtual Assistants (VA) and Bots as legitimate customer contact channels.

VAs have been around for a while, and have been given a huge boost both from a technological standpoint, but also market awareness due to the marketing efforts of some big players in the industry including Microsoft, Google and Apple. While I’m continually delighted that speech technologies are now main stream, I am still amazed at the considerable amount of attention VAs, and bots of all kinds (chat, messaging, etc.) are getting. I have seen numbers thrown about of tens of thousands of Bots that have, or are being developed on myriad platforms to automate processes and interactions. Truly, this digital workforce is changing the way we deliver customer service.

However, on the Experts Panel at Noble Systems User’s Conference in April, I stood on my soapbox again to reiterate a thought I’ve espoused for years, which I’m also hearing echoed in meetings with solution providers.  That is, you are only as good as your worst channel. While I’ve typically aimed this thought at IVR (which I still love and discuss below), bots and VAs give new fodder for my cause. Creating these digital workers, if not fit for purpose, are poorly designed or don’t deliver as expected, is a waste of time and a drain on the Customer Experience (CX).

This was one of many themes that weaved throughout the next three conferences I attended; Cisco’s customer contact analyst day, NICE Interactions/inContact user group (ICUC), and Genesys CX17 Indy. VAs and Bots were hot topics at all of these, but thankfully, all of the presentations had the overtone of cautioning people to carefully deploy them in places with impact, designed with specific applications in mind rather than open-ended “How can I help you” applications.  Additionally, all of these vendors were clear in voicing the opinion that platforms that support VA applications should be open, with APIs that allow for the ingestion of third party resources to broaden the capabilities of these assistants as well.

Is your company marching to an omnichannel strategy that includes VAs, but you are uncomfortable with the idea, or lacking the resources to develop this new channel? There are plenty of resources to help. In addition to the traditional contact center players, there are numerous companies in this space that cater specifically to the creation of digital agents – many who have been in the business of VAs for a decade or more than the press has covered the topic. Companies such as Creative Virtual, Interactions, noHold, Nuance, Omilia, SoundHound Inc., and others, have been delivering solutions on their own as well as through partners in the contact center space.

For instance, in May noHold announced it has added two Bots to Cisco’s Spark Depot. The Spark Depot is a place for users to browse through and add “integrations and bots that can accelerate business outcomes” to their own implementation of Spark. The  noHold Bots include a SparkHelp Bot for users to find answers about Cisco Spark, and a WebEx Bot that can answer common “how to” questions about WebEx.

A good example of a VA in action can be seen at the Royal Bank of Canada (RBC); Canada’s largest bank and the sixth largest retail bank by assets in North America. The use case for RBC was presented at April, at SpeechTek 2017 in Washington DC, by David Kapauan, Distinguished Architect, Contact Centre Technology at RBC. RBC’s contact center fields over 100 million calls a year and 55 million IVR self-serve transactions. It also supports 5.6 million proactive customer contacts as well. As part of its overall omnichannel strategy, RBC wanted to transform the DTMF-driven IVR with natural language speech technologies to improve CX, increase accuracy and reduce reliance on live agents.  To emphasize the spirit behind this strategy, David put up this quote from RBC’s CEO, Dave McKay, “In this increasingly instant and digital world, consumers don’t plan their finances as they once did, which means that we need to know our customer’s needs, their wants, their ambitions, and their fears, often without asking directly. We need to place ourselves in the context of their lives. In short, we need to think of ourselves as their financial Siri”.

RBC turned to Omilia and its omni-channel Conversational Self-Service platform, DiaManT to automate its credit card service line. RBC first piloted the Omilia solution in August of 2016, and got the green light to go fully live in February 2017.  The application handles unrestricted natural language and classifies customer requests into more than 130 intent categories and subcategories, while delivering end-to-end conversational self-service.  During the pilot the results were tracked bi-weekly for authentication in the IVR, self-service completion, agent-to-agent transfers, semantic accuracy and task completion rates (TCR). Throughout the trial, the application was also continuously tuned.  After just three months, the results were already pretty spectacular. The system delivered:

  • 93% semantic accuracy (the extent to which the system accurately extracted the correct meaning out of caller utterances)
  • 86% task completion
  • an increase in authentication from 44% to 71.5%
  • a world-class word error rate of just 5.6%
  • improved customer experience
  • increased call capacity
  • 59% reduction in IVR call abandonment rate over the DTMF IVR
  • less than 1% negative customer feedback on the new system

Plus these results are continuing to improve over time with more monitoring and tuning.

If you are crafting or delivering on an omnichannel strategy, VAs and Bots are a must investigate item. Don’t give up.  Deliver.  Voice is the most natural user-interface in the world, and with consumer acceptance of speech technologies at an all-time high, and consumers happy to self-serve, now is the time to elevate your customer service by supplementing with a digital workforce.

Category : Customer Contact

Nancy Jamison


Nancy Jamison is a Principal Analyst within Customer Contact within the Digital Transformation group at Frost & Sullivan. She covers all aspects of customer contact including cloud and premise-based systems and applications in the core areas of inbound/outbound routing, IVR, Workforce Optimization, and recording and analytics, with a particular focus on peripheral and emerging areas that impact the Customer Experience. These include speech technologies, omni-channel customer care, Big Data, digital marketing, Back Office Workforce Optimization, and Support Interaction Optimization.

Add Pingback
blog comments powered by Disqus