NETTING IT OUT
Virtual assisted-service is the use of technology to simulate the activities
of your customer service agents to help customers. From your customers’ perspective,
it’s assisted service because an (virtual) agent helps them answer questions
and solve problems. From your perspective, it’s self-service because
the agents on your customer service staff are not involved.
Human Digital Assistant (HDA) is a virtual assisted-service offering from H-care,
a software startup based in Treviso, Italy. HDA uses customer behavior on
self-service systems to generate events that trigger personalized, rules-based
responses in video, speech, and content in real time. HDA is quite lifelike.
Its implementation is customized to your requirements, and its delivery of
customer service is personalized to your customers. H-care introduced HDA
in September 2005. To date, five customer accounts have acquired the product.
On the PSGroup Report Card for Customer Self-Service, which we’ve modified
for virtual assisted-service, HDA exceeds requirements in its approach to virtual
assisted-service and in content management. The product needs improvement in
architecture, where packaged and higher-level integration facilities would
help implementation. The small customer base and H-care’s start-up characteristics
raise some viability concerns.
HDA is a technology breakthrough in virtual assisted-service. Your HDA implementation
can deliver a personalized customer service experience that integrates seamlessly
with your Web self-service systems. Its lifelike virtual agents will delight
your customers and help them do their business with you.
VIRTUAL ASSISTED-SERVICE
Software Simulates Customer Service Agents
In our customer service research and consulting practice, we talk about cross-channel,
cross-lifecycle customer service. By cross-lifecycle, we mean that customers
want and need your help at every phase of their lifecycles, from their initial
contact with you through their retirement. By cross-channel, we’ve
meant that customers want your help on every channel through which they interact
with you—the Web, kiosks, mobile telephones, and email for self-service,
your contact center, stores, and your field service force for assisted service.
New technology, most commonly called virtual agent technology, is causing us
to extend our definition of customer service. Virtual agent technology uses
software to simulate the activities of your customer service representatives,
your agents, helping customers. From your customers’ perspective, virtual
agents deliver assisted-service. Through integration with your self-service
systems, virtual agents help your customers answer questions and solve problems
when customers have difficulty getting answers and solving problems using
your self-service facilities. From your perspective, virtual agents deliver
another level of customer self-service—self-service because virtual
agents are your implementation of software, not your assignment of staff.
Given our customer focus, we’ll take the customers’ perspective
and call the customer service delivered by virtual agents virtual assisted-service.
Virtual Assisted-Service Is 40 Years Old!
Virtual assisted-service is not new. In fact, virtual agents are more than
40 years old. “Eliza” was the first virtual agent that I saw.
Well, ok, I didn’t see her. Forty years ago online meant teletyped
text. I typed a question. Eliza typed back an answer. And, ok, Eliza wasn’t
exactly a virtual agent. She was a virtual Rogerian therapist. More specifically,
Joseph Weizenbaum, now a professor emeritus of computer science at MIT, designed
and developed Eliza in 1966. Eliza parodied a real Rogerian therapist by
rephrasing “patient’s” statements and presenting them back
to the patient as questions. For example, the response to "My head hurts" might
be "Why do you say your head hurts?" The response to "My mother
hates me" might be "Who else in your family hates you?" Eliza
was named after Eliza Doolittle, the character in George Bernard Shaw’s
play, Pygmalion. You can still get virtual assisted-service from Eliza. Just
follow this link: http://www-ai.ijs.si/eliza/eliza.html.
More recently, in the past two or three years, virtual agents have been introduced
into the self-service context to help customers find information about products
and services and help answer their questions. Of this most current generation
of virtual assistants, “Anna” of Ikea is one of the most well
known. Anna is a cartoon character of a contact center agent. Her image is
displayed in a window that also has a search box. Type in a question and
Anna will “answer” it using content from Ikea’s Web site.
Illustration 1 shows Anna’s window.
ANNA OF IKEA
© 2007 Ikea
Illustration 1. Anna is a virtual agent that answers customers’ questions
by displaying content from Ikea’s Web site.
Anna and virtual agents like her deliver useful and valuable virtual assisted-service.
She’s much “friendlier” and easier to use than a plain
search box. She gives customers the feeling that they are getting real service,
not from the Web site and its typical service mechanisms, but from an agent,
a source with more knowledge about the site than they have. But Anna is simply
an attractive UI, an avatar, on a natural language processing search engine.
Eliza and Anna give us a view of the potential for virtual assisted-service.
With faster computers, huge gains in Internet speed and bandwidth, and advances
in video and audio synthesis, we can envision more life-like, animated virtual
agents who speak to us, listen to us, and, most significantly, help us get
our work done. Human Digital Assistant from H-care fulfills much of that
potential.
HUMAN DIGITAL ASSISTANT
Lifelike, Personalized Video and Speech
Human Digital Assistant (HDA) is the virtual agent product offering from H-care,
a privately held software supplier based in Treviso, Italy. H-care designed
HDA to be human-like in appearance, physical characteristics, and speech.
Further, it is designed to deliver personalized customer service by reacting
to customer behavior within self-service systems through events and business
rules to help customers by speaking to them, gesturing to them, and presenting
text and graphics to them.
Illustration 2 shows a real-world deployment of HDA at Telecom Italia. In
Illustration 2, the “Assistente Digitale” is a virtual agent ready
to help customers manage their accounts. The illustration is just a snapshot
of this
virtual agent. Go to the site, www.telecomitalia.it,
and follow “Assistente
Digitale” link under “Assistentza” at the bottom left to
see and hear how she can help customers.
ASSISTENTE DIGITALE AT TELECOM ITALIA
Illustration 2. This illustration shows Telecom Italia’s deployment
of HDA within the account management section of its self-service Web site.
By way of product background, HDA v1.0, the initial release, was introduced
in September 2005. HDA v2.2, the current version, was introduced in May 2007.
The next release, HDA v3.0, is planned for February 2008. To date, H-care
claims that HDA has an installed base of five customer accounts.
EVALUATING VIRTUAL ASSISTED-SERVICE PRODUCTS AND SERVICES
You should consider adding virtual assisted-service to your customer-self-service
applications to give your customers an alternative to picking up the telephone
when they can’t find answers or solve problems. A well-implemented
virtual assisted-service system lets customers continue to do their business
with you within the self-service context and gives you a lower cost-to-serve
alternative than contact center service.
So we’ve extended our framework for cross-channel, cross-lifecycle customer
service to enable the evaluation and comparison of virtual assisted-service
products and services. The top-level criteria of this evaluation framework
are listed and described below. The top-level criteria and sub-criteria are
shown graphically in Illustration 3.
•
Approach. Within approach, we describe and analyze virtual agent technology,
the video and audio content consumed by the virtual agent content, the integration
with Web self-service applications, customization, and personalization capabilities,
customer service functionality supported and not supported, and your work to
implement the product.
•
Content. Typically, you must create and manage the content that the virtual
agent delivers via video, audio, and text and graphics. We evaluate virtual
agent content against the same criteria that we have evaluated content for
cross-channel, cross-lifecycle customer service—content approach, content
model, content management services, metadata, globalization/localization, and
samples and templates.
•
Analytic functionality. You need to be able to measure, analyze,
and refine the customer service that your virtual agents deliver. Within analytic
functionality,
we examine and evaluate how a virtual assisted-service product collects behavior
and performance information and helps you in the analysis of that information.
•
Architecture. In architecture, we examine the implementation of a virtual assisted-service
product in order to evaluate how easily it can be integrated with your existing
customer service environment.
•
Product viability. This evaluation criteria allows us to assess the business
and risk in implementing a virtual assisted-service product.
•
Company viability. Where product viability examines product-oriented risk factors,
company viability examines risk factors with the product’s supplier.
PSGroup Evaluation Framework for Virtual Assisted-Service
