Characteristics
In addition to intelligence levels there are additional characteristics that are used to define intelligent agents:
-
Autonomy. The agent exercises exclusive control over its own actions and state without control of some external software entity. Autonomy implies the following (Turban):
- Goal-orientation. Accepts high-level requests indicating what a human wants and is responsible for deciding how and where to satisfy the requests.
- Collaboration. It can modify requests, ask clarification questions or even refuse to satisfy certain requests
- Flexibility. Action are not scripted, the agent is able to dynamically choose which actions to invoke and in what sequence, in response to the state of its external environment
- Self-starting. Unlike standard programs directly invoked by a user, an agent can sense changes in its environment and decide when to act.
-
Temporal continuity. Idea of leaving the agent to work once the task has been set.
-
Reactivity means sensing or perceiving changes in the environment and responding though actions.
-
Communication ability. Even the most basic software agent has the ability to communicate with other entities (humans, software programs or other agents) giving them various forms of social ability. In certain circumstances communication is necessary in order to delegate
Mobile agents v. resident agents
Resident agents stay in the computer or system and perform their task there. Mobile agents can move to other systems, transporting themselves across different system architectures and platforms. Examples of these software applications are agents with the ability to watch auctions and stocks and report back to the user if an item on auction or a stock reaches a certain price level.
Mobile agents add a third dimension in Fingar definition of agent software, which is mobility.
Turban describe as true intelligent agents only the agents that can be define in the three dimensional space of agency, intelligence and mobility.
The ultimate intelligent agents are agents that can reason under uncertainty, goal driven agents that can plan and learn by travelling across the network (Fingar).
Applications of intelligent agent technology
Intelligent agent technology is so interesting in the E-commerce universe as it provides a whole new approach to information handling, customer relationship management (CRM), advertising and marketing.
The main value offered by this technology is the possibility to locate and navigate the information overload created by tools such as data mining and log files, and also by analysing the data, extract meaningful trends and patterns that can be acted upon by the human end user. The process obviously saves time and the data extracted are of an infinite superior quality to any results that could be achieved by a human being in the same amount of time.
“The competent user’s decision is enhanced with information rather than paralysed by too much input”(Turban, p.2)
The following are some of the possible applications of intelligent software:
-
Personal assistant, in routine personal tasks such as grocery shopping or holiday booking. It is expected that intelligent agents for instance will be able to place an order for food shopping by collecting information from the favourite shopping list, from the fridge, in what is defined pervasive computing, the bank account and the diary to plan for the delivery of the goods.
- Search, retrieval and cost comparison, by searching in millions of databases and delivering the answer back to the end user when the task is completed. True intelligent agents will be able to prioritise what are the qualities and standards expected by the end user. Sometimes the cheapest product in not the best possible choice. The question is what is the product that offers the best possible value? This question is part of our buying decision making process and it will probably still take time a long time for agents to reach such a level.
-
Office activities, such as sales and customer support. Nielander, P (2005) offers an example of intelligent agents used in the Insurance Industry. It describes how the Fourth Generation System uses intelligent agent technology to complete 90% of the underwriting process without any human involvement. It requires the input of very little information from an applicant and leaves the work of finding out all the hard questions on the insurance application to the software (e.g. when a property was built or when the last penalty was issued). The information collected is extremely reliable as the software agent is able to access several other databases such as government or local council. The profile created is than used to decide whether the applicant will be accepted by the insurance company and/or into which rate tier. Benefits of the systems are the reduce costs, which means competitive prices and better profits.
- Domain experts. Intelligent agents could one day substitute costly expertise of real-world agents, such as translators, lawyers, diplomats, stockbrokers etc. and make it widely available.
How intelligent agents are used by Amazon.com and the online music industry
Amazon is a Website that teaches many lessons when it comes to using intelligent agent technology effectively.
The success of the idea behind Amazon is to provide the possibility to browse into the ‘earth’s biggest bookstore’, with a fantastic customer service and features that no physical store could ever match, such as book reviews, author interviews and personalised book recommendations. All this would have not been possible without intelligent agent technology.
Amazon uses software agents to:
- Search for the latest novelty in all categories (music, books, electrical etc.) and evaluate the best price, by comparison with the competition, so that the virtual store can be always kept updated with the best products and offers.
- Interpret marketing research information to support decision making
- Monitor and spot trends in consumer behaviour to adapt swiftly to changes
- Notify consumer when new CDs/books in their interest are available
- Provide consumers with personalised advertising
Their competitive advantage is the information and knowledge-based information system, the rest of the business in the Porter’s industry value chain is outsourced. Their system is their business. The software agent technology has been able to provide the tool to shift from mass production to mass customisation and has guaranteed the success of the entire operation.
In the process of giving the customers what they want, when, where and how they want it, intelligent agent technology plays an incredible part and an ever increasing number of organizations, including HMV, Reel and Tower Records, are becoming aware of the necessity of introducing these type of software in their system in order to remain competitive in the today demanding market.
Privacy
Intelligent agent technology brings up some concerns about the privacy of personal information.
Music websites offer recommendations to their customers based on profiles built by intelligent agents on the basis of demographic, geographic and psychographics data collected without the users’ knowledge. Recommendations are then generated by comparing profiles of customers with similar characteristics.
Another point that creates concerns is that agents can track movements on the network and in certain instances predict where the user is likely to go. Is this information supposed to be stored and whose information is it?
The debate on the ethics behind intelligent agent technology might go on for years. At present organisations are aware of the concerns and are thinking of possible solutions to the problem. Various standard have been proposed including giving the individual control on accepting the key principle of the technology.
Conclusions
Intelligent agent technology plays an important part in the success of E-Commerce, by allowing organisation to concentrate on the individual customer, leaving more time consuming chores to the software agents. The results are improved customer service, reduced costs and better profits.
End users are expects a certain level of service and automation when browsing in the network and are becoming familiar and acceptive of the technology amidst concerns about personal privacy. Successful organisations as Amazon and the expectations of consumers are pushing the industry to adapt and develop the technology, which is the only alternative to maintain a degree of competitiveness in today’s demanding market.
References
Nielander, P. (2005). Artificial Evolution. Best Reviews, October 2005, 100-102.
Turban, E. King, D. Lee, J. Viehland, D. (2004) Software (Intelligent) Agents. In: Electronic Commerce, A managerial Perspective 2004. Upper Saddle River, New Jersey: Pearson Education Inc.
Fingar, P. (1998) A CEO’s guide to E-commerce Using Object-Oriented Intelligent Agent Technology. Commerce net research Report access from http://home1.gte.net/pfingar/eba.htm [December 2005]
Fingar, P. Farhoodi, F. (1997) Competing for the Future with Intelligent Agents. In: Distributed Object Computing, “DOC” Magazine. access from http://home1.gte.net/pfingar/agents_doc_rev4.htm [December 2005]
A mathematical theory that studies interaction among self-interested agents
