INTELLIGENT AGENTS AND THEIR
APPLICATIONS IN E-BUSINESS
INTELLIGENT AGENTS
Intelligent agents are computer programs
that help the user to conduct routine
tasks, search and retrieve information,
support decision making and act as a
domain expert.
 Types of agents



Buyer agent
Seller agent
Need for agents
Information overload
 Automate repetitive tasks
 Aid in Decision Making

Six stages of sale transaction( Consumer
Behaviour model)






Identification: defining the needs to be accomplished by a
good, the constraints reducing opportunities and the
evaluation criteria to guide the selection process.
Product Brokering: searching for good that
fulfils the needs.
Merchant Brokering: gathering information
from the merchant to augment the evaluation process.
Negotiation: applying selected negotiation
strategies to a transaction.
Final Sale: transferring product/service
ownership.
Evaluation: evaluating the transaction
process (successful or not).
Application of Intelligent Agent…

Information Search and Filtering

Product and Vendor Finding

Negotiation Agents

Customer Service
Information Search and Filtering

Intelligent agent can help to determine what to buy to
satisfy specific need and where to buy…

Example:- Firefly Software(Barner &Noble,My Yahoo):
Firefly is a music and film recommendation system on the
web which uses intelligent agents to build up a complex
profile of user preferences(word-of-mouth advertising
principle): (Firefly was acquired
Yahoo uses firefly for website recommendations
It is a collaboration agent that makes recommendations
based on group evaluation of products


Product and Vendor Finding…






Intelligent agents help consumers decide where to buy by
comparing merchants offers….
Example:-Personalogic, Bargainfinder
Bargainfinder created by anderson consulting which compares
product prices of several merchants(best prices for cds)
The comparison agent takes the user product query and submits
in parallel to a group of online vendors by filling out the form at
each site.
The agent collects the query results parses the results and filters
of header, footer, ads to find vendor price information.
Problems- some sites do not want agents to extract information
only based on price
As agents themeselves are evolving, agent user may not be
willing to fully rely on the agents capability (such promotions and
offers, best offer with slightly higher price)
Product and vendor finding

Solutions:


Cooperative agent/vendor model: Vendors put the
product information that they would like to share with
agents in standard form such as Identify markup
language(IDML) which specifies how they would be
identified for search agents
Machine Learning Approach(In shopbot)


Learns how to shop without any cooperation from vendors,
can read HTML pages with product information
Jango( NetBot/ Excite) Jango , developed by
Netbot, Inc. where a user enters the name of a
product he is looking for and Jango automatically
determines which stores and information sites
are relevant. It is deployed at excite shopping
site.(www.shopping.excite.com)
Agents in Auctions
Ecommerce has brought the “too-manyoptions” problem to consumers.
 In e-auction, apart from finding the
product, negotiating for the product with
no fixed price to ensure that you win the
bid is a challenging task with limited
amount of time
 E-Auction is indirect type of sale and
buyers compete to get the right to buy the
product

Characteristics of Agents in Auctions

AutonomyEvery agent must act autonomously; that is, without
the user’s interference.

Proactivity The agents are configured to accomplish tasks based
only on their search plans and strategies of negotiation.

Reactivity The agents receive and process information sent by
other components; They wait for other user’s actions to decide
which action to take;They build up their thinking based on the
result of the processing made.

Sociability Every agent interacts with the auction sites to search
for information and sell or buy products; Some agents
communicate with each other.

Mobility The agents move themselves to the auction sites to
search for information and to sell or buy products.
Negotiation agents

As we know Prepurchasing phase may involve
negotiation in which price and other terms of
transaction are determined….
Example:- Buyer Agents like AuctionBot ,Tete-aTete (T@T), Kasbah

Tete-a-Tete (T@T)

Agent that uses co-operative negogiation approach to
retail sales
Negotiates based on several terms such as warranties,
delivery times, service contracts, return policies, loan
options, gift services and other merchant value added
services
Uses bilateral, multi agent negotiation environment


Negotiation Agents…..Auction Bot



Deployed at internet auction server at University
of Michigan)
A seller can create a new auction to sell products
by choosing from a selection of auction types and
then specifying clearing times, method for
resolving bidding ties and set up a reservation
price. The seller agent then manages the bidding
resolution.
It has API for users to create their own agents
and encode their bidding strategies and then the
agent bids and participates in the auction as per
user specification.
Buyer agent function





Registration: The agent must register in an auction site
autonomously, using its own identification.
· Product Search and Selection: It must be able to choose the
best available goods at existent e-auctions and search for
products that best satisfy its owner’s needs. The choice is made
considering existing constraints and criteria given by the user and
formalized by the agent generator.
· Negotiation Strategy Selection: The agent must be able to
plan the set of actions to win a bid based on its knowledge about
the specific type of auction. It uses a linear utility function that
varies as it reaches the bidding closing time. The agent may have
different strategies for each auction it gets enrolled.
· Bids: The agent must be able to make real offer for several
products and close the buying deal.
Auction Synchronization: Besides the individual strategy used
to act in an auction, the agent must be able to handle multiple
auctions offering its owner the best global deal
Vendor agent Function

Registration: The agent must register in an auction site autonomously,
using its identification.

· Site Search and Choice: It should choose satisfactory sites to offer its
products. This search is similar to the one made by the Buyer Agent visits
several auction sites to find out the best ones (lower fees, great variety of
products, etc.).

· Product Information Delivery: The agent is responsible for providing
information on the product when requested by the auctioneer or the
auction site. The agent must be able tostore and organize the information
about its products making it available when necessary.

· Monitoring: The agent can monitor current auctions it is enrolled to
verify the bids proposed and decide when to close the deal.

· Action Synchronization: it concerns handling the multiple places a
vendor exposes its product for sale
Customer Service



Several intelligent agent enhance customer
service by helping search and by providing help
to shopper
Notification agent: informs users of significant
events such as content change in a particular
webpage, user specified reminders.
Can be server based programs that keep
monitoring user specified sites