[Updated August 20, 2018]

In December 2013, I compiled this list of  trends to watch when thinking of Intelligent Assistants in the coming year. First and foremost was “Ubiquitous Natural Language Understanding” which, along the popularization of “Passive Authentication” and “Personalization [Driving] Innovation” defined how “Customers Take Control… Effortlessly.”

During the intervening years, each of these trends have built momentum; however, each at its own pace. Support of natural language by the likes of Siri, Google Now, Alexa on Amazon Echo, as well as hundreds of Web-based Enterprise Intelligent Assistants (EIAs). Thanks to the efforts of a dozen or so EIA platform and tools providers, those assistants are morphing into subject matter experts and advisers capable of carrying on conversations and providing personalized advice.

Meanwhile, as documented in posts on the Opus Research Web site, tremendous progress has been made in authenticating customers quickly and, in an increasing number of cases, passively. Following the example of Barclays Wealth in 2013, we’ve seen “conversational authentication” introduced by Investec in South Africa, RBC with 18 million customers worldwide and, most recently, Citicorp which claims over 200 million “customer accounts” around the world (roughly 250,000 of whom have enrolled voiceprints).

Justifying the acquisition and deployment of platforms that make things effortless for authenticated customers is not an effortless process for enterprise professionals. When Opus Research conducted a survey to discover the “Factors That Influence Enterprise Customer Experience Initiatives” we found that decision making most often fell to “working groups” that span IT, marketing, contact center, security and, ultimately, customer experience specialists. Financial Service companies have rushed to the fore for a couple reasons. Project champions build their business cases based on fraud loss reduction (thanks to strong, simple authentication) and then discuss the “soft” benefits that result from improved customer experience, including customer retention (or churn reduction) and increased sales revenue.

That said, our informal census of Intelligent Assistant implementations around the world spans multiple vertical industries. Telecommunications service providers, travel and hospitality specialists, consumer electronics companies, retailers and healthcare providers have played up the role of EIAs as resources that their customers can count on to help complete their tasks. The most forward looking anticipate a world where mobile or home-based personal assistants (MPAs) like Siri or Alexa will be able to talk to their EIAs to help their owners accomplish tasks quickly and with a minimum of friction.

Choice Versus False Choices

In the list above, I noted that “short- and long-form dialogues will co-exist.” It seemed obvious that, when users are given control, some will continue to populate search boxes with search terms (E.g. “train schedule” or “movie times”) and be happy with the initial results while others will grow more accustomed to full sentences such as “book me a seat on the next train from New York Penn Station to Baltimore.” If decision makers feel they must choose between providing “Answers” as opposed to “Conversations,” they needn’t. It is a false choice, and it is important to provide both.

As we enter 2016 there are a number of instances of similarly false choices that should not slow the technology adoption process. Here are five examples:

1. Automation versus Assistance: Self-service is seen by some as a “job killer” because it is simply automating processes that contact center agents used to perform. Yet our research on sequencing the omnichannel customer conversation shows that individuals routinely turn to over four different sources for information to support a transaction, including search engines, corporate Web sites, social networks and live contact center agents. Talking to an agent over the phone is included in almost 80% of those conversations. They turn to agents when the tasks get interesting.
2. Replacement versus Enhancement: Intelligent Assistants don’t replace existing IVRs or other Contact Center infrastructure. In the best case they leverage existing “scripts” for agent screen-pops, Web Q&A or speech-enabled IVR. They are, more accurately, the intelligent front end or “Smart UI” for existing knowledge management, business intelligence and speech analytics resources. Admittedly, some assembly will be required, but solutions will be the product of integration, rather than replacement of existing resources.
3. Avatars versus Assistants: If we’ve learned anything from the success of Google Now and its presentation of answers or suggestions via visual “cards” or spoken words, it is that there is no explicit need for a named entity like Nina, Alexa or Siri. Yet we can also observe that a vast number of end users like it when their automated assistant has a visual manifestation when on a Web site or a personality over the phone. It is an evergreen topic that should not slow-down the broad implementation of Enterprise Intelligent Assistants as the resource that serves as the first contact between customers and the companies with which they want to carry out business
4. Answers versus Conversations: A successful customer experience is often equated with rapid resolution of an issue or completion of a transaction. By contrast, successful customer relationships are more like conversations carried out over time and with continuity and context awareness. Firms that have focused on the former have developed Answerbots, which are a fine category for single-purpose, one-and-done chatbots. Yet, after gaining experience with these short-form experts, many are exploring the use of NLP, ML and analytics to provide relevant data to inform the more complex conversation. It is not one or the other.
5. Speech versus Text: For a long-time speech zealot, like myself, this has been one of the hardest reality to accept. Upon reflection, it is obvious that speech recognition is not the goal, any more than speaker recognition for the purpose of authentication is. Natural Language Understanding has been the real catalyst. Enabling individuals to communicate in their own words – whether they are talking, texting or otherwise messaging with an EIA or MPA – is the real confidence builder. Don’t choose between speech recognition or support of text. Do both.

When it comes to Intelligent Assistants in 2016, heed the directions attributed to the late Yogi Berra, “When you come to a fork in the road, take it.”

Categories: Conversational Intelligence, Intelligent Assistants, Intelligent Authentication

Tags: Contact Center, customer experience, Natural language processing, intelligent Assistants, Mobile CX