Virtual Agent Technology - Essay Example

Topic: Virtual Agent Technology Leah Cocolicchio Tony van Krieken ITM102 Business and Information Systems February 14th 2007 Contents Contents 2 Executive summary 3 Introduction 4 Explanation 5 Historical background 5 Who uses the technology 5 How do you keep the virtual rep experience from feeling impersonal 5 Handling Off topic questions 6 Technical issues 6 Managerial issues 7 Case study 8 A Success story 8 Various Agents In Use 8 Concerns about Agent Use 10 Security and Privacy 10 Trust and Reliability 10 Conclusion 11 References 13 Executive summary The purpose of this dissertation is to announce the success of the virtual agents, in its wide acceptability in the customer service environments with consistency in messaging. It calls back to the live agents, humans, who handled these services in servicing the customers for their queries and faced many psychological problems relating to the human emotions and time factors. Technological advancements have come up with such skyrocketing techniques that can formulate the speech and can understand nuances of human languages. Thus it brought about Virtual Agents. The data collection is primarily from the internet sources and e-journals. The limitations are the various technologies that are required, to set this up in any organisation. The idea is to make an organisation belief that such technology is of great importance concerning their requirement for customer services agents. The primary investment which needs to be made by the organisation to setup the concept of Virtual agents to handle customer queries and the trust in the technology to yield better revenue and service. The research is done on three key enterprises RBC (Can.), Dr. Frank Inc. (U.S.), and Sutton and East Surrey Water Co. (U.K). The primary conclusion to this research topic is that the success of this technology is not in research laboratories but in the hearts and minds of people and corporate institutions that will be using it for their well being. The ability to trust it over time and enveloping the merits and demerits of the technology is the ultimate success of it. I would recommend this technology to be used in place of live customer agents in an organization where decisions taken ranges from simple to complex. Introduction Increasing the flexibility of I.T. systems is a common strategy to cope with ever more complex and demanding user requirements. Two common approaches to creating flexible, loosely-coupled systems are asynchronous message-passing architectures and synchronous remote procedure call (RPC) architectures. Software agents, one strand of research and development into flexible I.T. systems, extend message passing architectures to exhibit properties of being social, reactive, proactive and autonomous (Wooldridge, 1995). The motivations of web service designers are similar to those of agent designers (Zimmermann, 2003), despite the differences in technology. Explanation Historical background Phone-based customer service, as necessary as it may be for a business, is expensive. An increasing number of companies are using the Internet as a customer-service channel, offering text chat with both automated responses and real-life agents. Who uses the technology Virtual service representatives basically serve as [phone] call avoidance. A virtual representative is a server-side program that functions like the frequently asked question page on the Web site, but it also serves as a tour guide for the site. It simulates a text chat session with a live agent, with page-push or co-browsing capability. How do you keep the virtual rep experience from feeling impersonal The standard content on a Web site tends to be dry. A little more life can be added with a picture of the virtual agent. Depending on the question asked, it can have a visage that greets the customer, one that smiles, frowns or looks confused. The virtual representative can also be conversational and answer off-topic questions. Depending on how the client wants to be answered those questions, we can give at a personality - it can be polite or very sassy. A common question for the bot seems to be, "Where do you live" The answer can be scripted is, "I live in a server in Company X's data center." Then people might ask, "What's the weather like there" We'll push a Web page displaying the weather at the company's city and say, "Here's the weather for Cincinnati, Ohio [my home town]. But I don't know what the weather is like anywhere else" (The Industry Standard). Handling Off topic questions People tend to get more off-topic with the software agents than with live chat, where a real person is on the other end. They're not used to asking a computer system a question and receiving seemingly intelligent responses. The customer is made aware of the fact that they are speaking with a virtual service representative with very precise answers. Technical issues The various technical issues relate to the technology in use to support the concept of virtual agents. The automated software's require to be designed for specific businesses and are not readily available. It must be designed well for the success of any business to cover all the domains available in that business. Managerial issues The commonest factor is that it must be of service for the business in the long run and yield profits and satisfaction for the customers. The most important factors are: 1. Goal-directed. The user or other invoking entity specifies a goal (desired world-state) rather than simply a task or "recipe" of tasks to be performed. The idea is that the software agent (coupled with a smart infrastructure) will figure out how to accomplish the goal rather than simply mechanically carrying out pre-programmed tasks. 2. Level and areas of intelligence intelligence. It's quite possible that a software agent could be quite intelligent in some respects or domains and quite 'dumb' in others. 3. Ability to Learn. A software agent could come with strictly pre-planned knowledge, or could acquire knowledge, expertise, and even some aspects of intelligence based on input from its environment. 4. Autonomy. A software agent would typically have at least some ability to work independently from the user or invoking entity, but the degree of autonomy could range from simple, immediate task execution (offering some value-added service) up to endlessly pursuing a goal once initiated. 5. Mobility. A software agent could move (migrate) around the network, either of its own volition or as moved by the agent infrastructure to balance or optimize network or application performance. Case study A Success story "Dr. Frank's telemarketing costs decreased from approximately $15-20 per order with live agents to approximately $4 per order. Average sales jumped to $41 per call-- $13 higher than per-call sales by live operators" (Winslow, 2006). The successful use of Virtual Agents in Sutton and East Surrey Water has saved the company "thousands of pounds" (Davies, 2006) and they have been awards three national and international awards-including "most effective use of IT in Utilities" (Davies 2006). Various Agents In Use The following table details sample agent applications either in current development or use: Agent Class Application Originator Description Deployment User Interface Open Sesame! Apple Learns & analyzes user behavior, suggests automation of specific functions. Commercial MS Agent Microsoft Simplifies complex tasks through natural interface, including speech recognition. Commercial Monitoring SUMPY Univ. Memphis Performs disk maintenance functions for Unix servers. Lab Information Retrieval WebCompass Symantec/ Quarterdeck Calls various public search engines, then carries out search & summarizes found content. Commercial HotBot Hotwired & Inktomi A public search engine. The agent resides in the searching system of home server. Commercial BargainFinder Andersen Consulting Searches several products matching user preferences based on price. Semi- commercial Information Filtering NewsHound Knight Ridder Personalized digital newspaper based on pre-determined preferences. Commercial PointCast Network PointCast Personalized information (push)delivery service. Commercial Notification URL-Finder Net Mind Periodically detects URL changes & then notifies user. Commercial AdHound Classified Warehouse Searches & tracks the changes in user- specified advertisements. Then notifies user by e-mail. Commercial Process Automation Edify Workforce Edify Customized, automated self-service agent for personal communication & activities. Commercial Collaboration Firefly Firefly (Microsoft) "Word-of-mouth" marketing. The agent suggests products based on prchases of users with similar profiles. Commercial Julia Chatterbot Carnegie Mellon Univ. Natural language chattering robot. Commercial (Other co.s) AuctionBot Univ. Michigan Creates & manages automated Internet auctions. Lab Tete-a-Tete MIT Media Lab Negotiates a deal based on price & other terms like warranty, payment method & shipping cost. Commercial (Frictionless) Decision making DSS Agent Micro Strategy Integrated query & reporting, analytic & decision-support workflow. Commercial General Use Ginko IBM Agent system development tools. Commercial Concordia Mitsubishi Electronics Mobile agent solutions. Commercial (MIT. See http://web.mit.edu/sirkin/ecom/final.html) Concerns about Agent Use Security and Privacy Software agents, just like the human kind, represent a principal with apparent legal authority and responsibility. Agent actions must rely on prior user authorization, and the rights of the agent cannot exceed those of the user. Trust and Reliability The business must increasingly trust the agent to "do the right thing" as autonomy increases. At issue are the relatedness between the user's intentions and the agent's actions (accuracy), the reliability and completeness of searched and filtered information (precision), and the ability to detect and ward off threats like viruses. The required training or learning process may itself threaten user acceptance of agents. Immature agents are liable to make mistakes, such as retrieving information from out-of-date or unreliable sources. If the loss in credibility during this stage is significant (so the costs of continued use appear greater than the benefits), the user may curtail or discontinue agent use. Conclusion Software agents are central technologies in shaping networked markets. But the technology's potential for increased efficiency and lower costs must be viewed with an eye to its potential for accidental misuse and intentional exploitation. Taken together, we can identify several trends we expect to develop or continue over the next several years. Transaction costs will continue to decrease. Agents make the custom retrieval and presentation of information less costly. The change in expenses and the availability of information to customers and vendors may change the balance of power in making transactions. Markets will either converge based on price or diverge based on custom offerings. The ability to differentiate products is a central factor in determining how a particular market develops. Agents' ability to investigate an entire market space will quicken the pace of the convergence to one regime or the other. Also, their ability to customize information from varied sources suggests that more products (especially information goods) will be differentiable. Trusted-agent certification authority will appear and gain importance. Whether the authority is centrally located or resides on local servers within a market depends on the scalability of the technology that is broadly adopted (MIT. See http://web.mit.edu/sirkin/ecom/final.html). The replacement by virtual agents is certain in the near future as technology evolves greater integration of knowledge and intelligence to handle and capture almost every resource. References Davies, J. (2006). Speak Easy.(public utilities make use of speech Recognition technology). Utility Week, np. MIT. See http://web.mit.edu/sirkin/ecom/final.html The Industry Standard. 'Talk to the bot'. See: http://www thestandard.com/article/0,1902,15756,00.html Winslow, D. (2007). Next Generation Speech Technology: maximizing your revenues. Response, 84, 12-14. Wooldridge M. & Jennings N. "Intelligent Agents: Theory and Practice. Knowledge Engineering Review. Vol. 10:2. 1995. pp. 115-152. Zimmermann O, Tomlinson M, Peuser S. Perspectives on Web Services Springer, 2003. Read More