Interactive designers aim at developing interactive systems that bring forth positive responses from the users (Rogers et al., 2007), including responses such as making the users feel motivated to play, learn, be social, and be creative.
Virtual agents are one of the interactive interface elements that interactive designers use to make the users feel emotions such as happiness, comfort, and a feeling of being at ease. They are also used to convey emotional states. Virtual agents are created to give users a sense of having a companion. It is assumed that interaction with the virtual agent would encourage the user to try things out and would make the user feel comfortable in using the application. These agents usually take the form of a friendly character such as a cute bunny or a pet dog.
One of the first virtual agents developed by Microsoft was Bob. However, it was never commercially released as users found it too childish and cute. Microsoft then came out with the virtual agent Clippy, a cartoon paper-clip that exhibited human-like characteristics and a warm personality. It was part of the environment of the Windows 98 operating system and was usually displayed at the bottom of the user’s screen. It appeared “whenever the system thought the user needed help carrying out a particular task” (Rogers et al., p. 189). It communicated with the user through text. Users could enter a query and Clippy would display the information that’s most closely associated to the user’s query. It was an interactive way of displaying the online help.
It was not well-received by the users, however, as users found it intrusive and trying. It distracted them from their work. Personally, the writer found it cute and fascinating at first. However, it became annoying and distracting in the long run, as it kept offering help even when the help was not needed. It also took up unnecessary space on the screen, which obstructed some of the applications that were currently running. As such, it wasn’t persuasive as it did not encourage the user to look up information about something or try out new things. In the end, it was rendered useless as the user would just turn this feature off. As Rogers et al (2007, p. 189) stated, negative emotional responses are elicited from users “when the appearance of an interface is too noisy, garish, gimmicky, or patronizing.”
The virtual assistant Anna of IKEA.com is somewhat different from Clippy in that she “seems” capable of answering back. Anna comes in the form of a cartoon-like woman that the user can communicate with through text, in a manner that’s similar to instant messaging or online chat.
IKEA portrays Anna as an online assistant that is supposed to be able to answer the users’ queries. However, although Anna does answer some of the simpler questions quite intelligently and proves to be helpful in some queries, she is generally not capable of answering the users’ more complicated questions about IKEA’s products, such as when the user’s questions are the sort that they should be asking a real live sales person.
While some users find Anna to be more of a greeter than an online assistant (Scott, 2008), the writer thinks that Anna does serve as an online assistant, but not in the context of helping users complete their shopping tasks or providing them with product information. Rather, Anna is more of an online assistant in the context of providing the users with online help, which consists mainly of information on how to use the site or how to navigate the site.
This supports the claims of experts who criticize the anthropomorphic approach. One such expert is Shnelderman (Rogers et al.) who claimed that anthropomorphic interfaces, such as those that involve screen characters or first-person dialogs, come out as being deceptive. They give users the notion that they can converse with the user in much the same manner that a real person would. This raises the users’ expectations of the virtual agents’ capabilities that the users just end up feeling frustrated and annoyed when they realize that the virtual agent is not capable of such things after all. Users then become disillusioned and immediately lose trust of the virtual agent. As a result, these types of interfaces are more counterproductive than productive.
Proponents of the anthropomorphic approach claim that the use of interactive systems that have human-like attributes and personalities make them more fun and enjoyable to interact with (Rogers et al.). They also assume that the anthropomorphic interfaces do a better job of motivating users to perform tasks such as purchasing goods online, compared to presenting the user with abstract and cold computer language. These interfaces also allow users to ask their questions in a more natural way, using plain English, instead of making them think of relevant keywords for their search.
However, while these may be true for some anthropomorphic systems, it definitely isn’t the case with the virtual agent of IKEA.com. Anna of IKEA.com just leads users to think that it possesses human levels of intelligence, only to later find out that it is just a bot that provides a series of pre-programmed answers to the common questions that users ask. As such, it does not do a good job of persuading the user to purchase products from the site or to even explore the site further. Instead, it can disappoint the user to the point that the user will just leave the site without having accomplished anything.
Rogers, Y., Sharp, H., & Preece, J. (2007). Chapter 5: Affective aspects. In Interaction design:
Beyond human-computer interaction (2nd ed.) (180-215). Hoboken, NJ: John Wiley & Sons
Scott, D. M. (2008, August 20). Anna from IKEA is intellectually challenged (but she has asense of humor). Retrieved from http://www.webinknow.com/2008/08/anna-from-ikea.html