Customer Service NLI Virtual Assistants — 22 November 2011
What makes an outstanding Online Virtual Assistant?

The massive strides forward in information technologies and in particular artificial intelligence has led to a paradigm shift in our expectations in human-machine interactions.

When users interact with a virtual assistant over the Internet or via their mobile devices, they expect to find it as natural, lively and encouraging as chatting with their friends on MSN or talking to their relatives on Skype. A glance at some of the conversations between real users and any of the world’s most popular web-based assistants would suffice to prove this. The question, however, is: Do virtual assistants really satisfy users’ needs?

Considering all the fuss over Apple’s Siri, Customer Service departments in companies all around the globe must be asking themselves whether a virtual assistant can help strengthen the relations between organizations and customers, ultimately providing an additional channel for boosting the level of customer satisfaction.

Finding a balance between services provided and customer satisfaction, challenging as it is, is not a matter of just being edgy. What factors make a service become “quality” service? Can you improve customer satisfaction by automating the services that you provide? How can a virtual assistant fulfill customers’ expectations?

These are all questions that can be answered from the perspective of some of the most long-time and enduring cognitive sciences, like non-verbal communication or linguistics, and most recently, from the advances in natural language interaction and speech technologies.

The truth is – and this remains a key factor – it is mostly the perceived personality of the virtual assistant that elicits different reactions from users, both positive and negative. Therefore, if properly applied, the use of personality cues in virtual assistants will provide opportunities to improve customer experience.

These are some of the most valued personality cues in a virtual assistant:

  • Small Talk: This term was originally coined by specialists in discourse analysis (the study of written, spoken or signed language) to refer to irrelevant conversations that help interlocutors (the person who is taking part in the conversation) to engage in social and personal binding. It is sometimes called “elevator talk”. Can you imagine why? Typical small talk topics include the weather, movies, sports, hobbies, activities, and so on. Small talk conversations can become the benchmark of a virtual assistant, both in positive and negative ways. So, it is extremely important to align corporate image and small talk when designing the personality of a virtual assistant.
  • Proactivity: The idea of proactive behaviors in human-machine interactions has been around for some considerable time already. When people engage in verbal interactions, they usually expect their interlocutors to move forward in the conversation. This is something that many virtual assistants would claim to do, but in reality only a few actually achieve.
  • Humor: It is usually the case that users rate virtual assistants as more competent when they exhibit humor. A number of experiments suggest that humor can have many positive effects in human-machine interaction.
  • Non-verbal cues (appearance, posture, gestures, facial expressions, etc.): Most virtual assistants are presented to the user as embodied agents themselves, with a given set of predefined gestures and facial expressions that, when emphasizing the information delivered graphically and/or verbally, grants a natural effect that is inherent to humans. In addition, it is an observed fact that the expression of emotions provokes positive empathy from users.
  • Speech Synthesis: What is a virtual assistant without a voice of its own? This is clearly one of the most relevant and human-like features for a virtual assistant. Different studies suggest that users tend to rate higher those human-machine interactions where the assistant provides spoken feedback. The use of Text-To-Speech technologies allows the possibility to take advantage of paralinguistic features such as tone, loudness, and so on, and make them coherent with the assistant’s personality.

Ultimately, it is always advisable to think about the automated customer service experience in terms of end users. Assessing overall performance of virtual assistants is not a matter of just answering users’ questions correctly; providing a personality to an assistant is a sure thing for eliciting positive reactions from them.

Related Articles

Share

About Author

Carmen Del Solar Valdés is a Knowledge Engineer at Artificial Solutions. From 2005 to 2011 she worked for the University of Seville's spin-off company Indisys. During this period, she was also a member of the Julietta Research Group at the university. Carmen holds a B.A. in English Language and Literature from the University of Seville and she spent the first year of her graduate studies as a visiting student at the Department of Linguistics at the University of Pennsylvania, USA.

(1) Reader Comment

  1. I thought this was very insightfully written. Especially humor can go a long way towards winning over the user. If a VA says something funny, people don’t seem to mind if the answer didn’t really fit the question.

Leave a Reply

Your email address will not be published. Required fields are marked *

*


× 7 = twenty one

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>