In the previous blog in this series, Tami Waggoner talked about looking at customer engagement through the client’s eyes, and how that shapes Accenture’s Applied Customer Engagement (ACE) offering.

Another key capability in the ACE repertoire is the use of modern virtual assistants. The maturation of the digital assistant capability has made it possible for the contact center to handle a much broader range of requests. When calls are “deflected” to virtual agents, the virtual agent now performs as well as a human and, in some cases, handles requests better than a human can.

ACE’s ability to deflect calls to the virtual agent rests on two key factors. First, customers are much more ready to use and trust virtual agents. There are now more than 100 million intelligent voice devices in the market place and we expect the number to rise given the appeal of their capabilities to customers.1 Second, the maturity of the underlying natural language understanding (NLU) and artificial intelligence (AI) technology – leveraging the cloud – is allowing providers to reach in some cases an 80 percent containment level within weeks rather than months, thanks in part to accelerators such as prebuilt intent libraries for dialogue flows and utterances.2

This is a big step forward from just a few years ago, when chatbots were first introduced. Neither the technology nor the customers were quite ready, and few if any organizations had the requisite digital capabilities in place. Now, digital-based features are mature enough to support big advances in contact center operations.

The second generation of virtual assistants goes far beyond the simple “frequently asked questions” capabilities of the original chatbots. For example, these assistants are now connected via application programming interfaces (APIs) to back-end systems, allowing them to address transactional (guided and relatively simple) enquiries as well as procedural inquiries, which may be complicated, with multiple dialogue flows bringing the customer back to the previous step before moving forward again.

We are still in the early stages of the virtual agent revolution. We expect the virtual agent to become customer’s permanent sidekick, with his or her best financial interests at heart. Expanded AI capabilities allow the virtual agent to become proactive, understanding the customer’s financial situation, anticipating needs and offering guidance.

We are mostly at the second-generation level for virtual assistants, following an evolutionary path that looks something like this:

  • 1st generation – Mostly chat involving semi-scripted interfaces such as frequently asked questions (FAQs);
  • 2nd generation – Chat and voice with a richer set of intents based on NLU including transactional enquiries with back-end connectivity through APIs;
  • 3rd generation – Richer NLU with more complex procedural intents with high level of containment in scope;
  • 4th generation – Personalized virtual agents adapting to customer’s style (both written and verbal) with underpinning prescriptive AI to guide financial needs;
  • 5th generation – Introduction of empathy into interactions (this is still “out there”).

Within a few years (at the 4Gen level) customers should simply ask their virtual assistant “What do I need to worry about today?” And that is only when their virtual assistant does not proactively contact them. This should be good for the efficiency and effectiveness of contact centers, but even better for the customers themselves.

In the next blog in this series, we will look at some of the talent issues related to the rapid evolution of the customer contact center. And, to learn more about ACE, watch our video below and read our report: Accelerating Growth with Applied Customer Engagement.

Video Transcript


  1. “Amazon says 100 million Alexa devices have been sold – What’s next?,” The Verge, January 4, 2019. Access at:
  2. Based upon Accenture’s experience and work in this area.

Submit a Comment

Your email address will not be published.