It is possible to distinguish at least four types of agents on the Internet:
- Agents for information retrieval;
- Monitoring agents;
- e-commerce agents;
Naturally, the first two agent categories are most relevant for undertaking strategic intelligence activities on the Internet. For this reason, we previously described them in some detail. In the optic of this work, the two remaining categories, although promised with an unquestionable future, are of less interest. We will therefore describe them in a brief way, while especially putting emphasis on the possible interactions for information retrieval and monitoring.
Agents for Commerce
The intelligent agents dedicated to commerce are in full expansion and are likely to revolutionize commercial practices in the years to come. The enormous interest in e-commerce has lead to the creation of many more or less advanced software applications. The problems are very similar to that of information retrieval. With an international, scattered, and plethoric offer, the consumer has more and more difficulty choosing the product they want. In a symmetrical way, faced with an ever more atomized and hard to satisfy demand, the commercial website does not always manage to guarantee a personalized offer adapted to the true needs of the consumer. Therefore, as Raphael Richard states, " in the field of commerce, the intelligent agentsí mission is to make the choice of shops, brand-names, and products easier for the consumer, based on their tastes, but also to allow the sales and marketing staff to better understand the demand ".
Two categories are therefore distinguished.
Agents controlled by the customers, often called shopping agents (or ShopBots). Their objective is to facilitate the purchasing process. Indeed, as for all other types of information retrieval, identifying and checking the quality of a commercial offer can be extremely difficult on the Internet. The traditional tools (search engines and directories) are quickly found to be ineffective when finding, checking, and comparing a commercial offer. Imagine that you are looking for a book. The traditional approach consists in trying to find some stores likely to offer the book, using sites such as Yahoo and AltaVista. Then, based on some keywords, you obtain a long list of Internet addresses that are probably not even on-line bookshops, but simple bibliographical references... And even if you obtain some bookshops, you have to go through their catalog to find the book, before trying to compare the various offers. In short, for a detailed search, you risk spending a whole day on just one book...
Search agents can speed this process up a little, but they are limited to a completely manual and linear approach. This is why shopping agents were introduced, which, once you indicate the name and or the author of the book you are looking for, connect to the Internet in a completely autonomous way, visit various on-line bookshops, identify the book, compare the different prices, and finally buy the least expensive one! Unlike search agents, e-commerce agents do not use keywords, but the names of products or trademarks directly.