Are Intelligent Virtual Assistants Ready for the Enterprise?

“The IVA will map what the users say to the enterprise terminology, putting even less demands on consumers to do what they want,” Jonsson adds.

While much of the discussion around enterprise IVAs center on their use by consumers, virtual assistants provide significant advantages for workers as well, who are using IVAs rather than help desks or Internet searches to find critical information.

“IVAs can help the workforce optimize its procedures by being able to come back with consistent information from different back-end systems and services,” Jonsson says. So field force workers, for example, can get answers to their queries, such as parts specifications, product availability, and other information, without needing to wait for a live agent at the help desk.

Some systems that also incorporate predictive analytics and machine learning will ask meaningful, relevant follow-up questions if they’re unable to answer an initial query without more information. These systems will ask one or more additional questions to elicit the needed information from the user, rather than simply saying: “I need additional information.”

While IVAs cannot handle some of the most complex queries, the systems enable enterprises to quickly collect meaningful information about the customer and his or her query, Jonsson adds. Enterprises can mine the collected data for better understanding of customers/workers and what they are trying to achieve.

“For this reason, at Artificial Solutions, we are focusing heavily on the analytics part of the IVA logs,” Jonsson says.

In the most advanced uses, these systems are becoming virtual advisers, going beyond simply answering a customer’s or employee’s initial query to incorporating artificial intelligence, machine learning, and predictive analytics, leaving live agents for only the most complex work.

A good example of this, according to Miller, is the use of IVAs at Hyatt Hotels, which has a system that knows when to best hand off a call to a live person for booking a room, upselling, and additional service.

“This leads to a better user experience,” Miller says.

Humans are removed from the ordering process almost entirely with the Domino’s Pizza Dom, a mobile IVA that uses machine learning and predictive analytics to know not only the basic customer information, easily determined from the mobile phone’s number, but also a customer’s typical order.

“For example, if you and your buddies get together every Thursday night and always make the same order, Dom will ask: “Hi, Dan, do you want two pepperoni pizzas like you order every Tuesday night?”

Domino’s has even gone the additional step of enabling customers who have downloaded and registered on the app to purchase a pizza via Twitter by sending an emoji, a capability the company touted during a 2015 television advertising campaign.

That campaign strives to play up one reason people use the IVA and emoji, and these systems in general: They’re fun. “People interact with our virtual assistant, Indigo, many times just for the fun of it,” Jonsson says. “That is the reason why we, last year, launched Elbot on mobile, so people have something that only chats.”

Using IVAs to Drive Sales

However, enterprises want to do more than simply entertain customers and workers with their IVAs.

“Retailers see IVAs as a customer-facing experience for their omnichannel efforts,” Miller says.

“Enterprises are looking at how they help customers and drive sales; IVAs serve as entry points to drive sales,” adds Jane Price, vice president of marketing for Interactions. “Businesses that interact with their customers are better at retaining them.”

“An IVA can do many of the same things as a customer service rep or sales agent; it becomes your personal sales agent,” says Phil Gray, executive vice president of business development at Interactions LLC.

IVAs provide better interactions than IVR systems, which are often difficult to use and inconsistent in their ability to handle customer needs, Price adds. IVAs also provide better consistency because they employ the same logic and intelligence, whether the customer or worker makes contact via a desktop, laptop, or a mobile device. IVRs and Internet-based systems, by contrast, employ different technologies, so use of one versus another will differ significantly.

“People interact surprisingly naturally with IVAs,” Jonsson says. “They follow social protocol with greetings and politeness. They even say thank you to the IVAs. You would not see that when people are doing a Google search. Not only do people interact more naturally with the IVAs, but they expect the IVAs to do so as well. They expect the IVA to be aware of the current dialogue [and apply it to available technology], such as the geolocation of the user; they expect it to follow them if they switch topics and to understand them if they go outside of the enterprise’s domain, like if they start to talk about the weather. They’d be able to express themselves with their own terminology and not have to learn specific words or commands.”