Information Technology I Software Agents

Information Technology I Information Technology I Business Intelligence II Software Agents Frictionless commerce??? • Low search costs • Strong price competition • Low margins Information Technology I Frictionless commerce??? • Empirical data suggests that is it still an elusive dream… – Amazon charges 20-30% higher prices than its online competitors but still manages to maintain a 85% market share • [Brynjolfsson and Smith 2000] – “Predictions that lower search costs would increase competition, forcing prices to fall to cost […] have not been realized. Average prices are well above cost and are flat or rising over the sample period.” • [Clay et. al. 2000] What is going on??? Information Technology I Amount of information is increasing Getting the right information is daunting Information Technology I Electronic commerce is still primarily a humanhuman-centered activity Buyer Seller Web Select product Inform prospects Perform transaction Perform transaction Receive product Fulfill order Request service Provide service Example: Use the Web to organize a trip • Where to go? • Where to stay? • How to fly? • What to see? • Where to eat? • Etc. etc. etc. Wouldn’t you rather use a travel agent??? Information Technology I Enter software agents • A software agent is an autonomous (software) actor which can take actions towards its goals • Software agents can help their human masters find information, make better decisions and obtain better transaction outcomes What can software agents do • Select one or more actions based on “rules” • Select actions based on knowledge about their users • Have dialog/negotiation with other software agents • Autonomously learn over time Information Technology I An early example: Intelligent email filtering agents Agents in the buy/sell process • What to buy? – Recommendation agents • Where to buy? – Price/merchant comparison agents • How to buy? – Automatic negotiation agents Information Technology I What to buy: Recommendation agents Example: Amazon Screenshot of recommendations page from www.amazon.com: "Welcome to Recommendations. Here are our recommendations for you." Information Technology I Collaborative filtering vs. personal agent approach • Collaborative filtering – Is based on forming clusters of “similar” customers who visit a given site – Personalization engine and data are “owned” by retailer/intermediary • Personal agents – Learn individual consumer’s preferences by trial and error by observing the consumer’s interactions with all sites – Are owned by the consumer How do agents learn? • Several approaches – Adaptive neural networks – Reinforcement learning – Genetic algorithms Information Technology I Adaptive Neural Networks Inputs: • Product attributes Output: • w13 w14 1 Probability of purchase 3 w36 w23 w15 4 w46 6 w24 2 w56 w25 5 Inputs Output Hidden Layer Adaptive Neural Networks • Start with rough guess • Each time, observe consumer’s response and use transaction as the next “training” example w13 w14 1 3 w36 w23 w15 4 w46 6 w24 2 w25 Inputs w56 5 Hidden Layer Output Information Technology I Reinforcement learning Agent Input Table W ( S, a ) Action Selector Action Environment Environment LookUp State Recognizer E Learner Reward Genetic Algorithm Case Study: Amalthea: Amalthea: A Personalized Information Discovery Agent Ecosystem Information Technology I (Screenshot of Amalthea: A Personalized Information Discovery Agent Ecosystem.) Amalthea architecture Information Technology I Amalthea functionality • Amalthea creates an ecosystem of agents, which search the web for “interesting” sites • Each agent searches for sites which contain a given set of keywords • Amalthea users rate the returned documents • Based on user ratings, agents evolve – “Worthless” agents get killed – “Useful” agents are allowed to “mate” (I.e. combine the keywords they are looking for) and form the next generation • Over time, this evolution process results in increasingly good fit with the user’s interests Information Technology I Amalthea genetic evolution Information Discovery Agents Information Filtering Agents Amalthea performance Information Technology I Benefits for customers • Reduce search time/effort • Make better recommendations • Improve over time • Tailored content and advertising • One-to-one marketing • Etc … Benefits for providers • Higher customer satisfaction • Higher loyalty – … because benefits increase over time • Accumulate useful data for “market research” – … but must be very careful with privacy laws!!! Information Technology I Where to buy: Price comparison “shopbots “shopbots”” Screenshot of search for Guinness World Records 2000 book at price comparison "shopbot" and the results from different online vendors. Information Technology I Limitations of current shopbots • Do not necessarily display the information the consumer really cares about • Do not capture the consumer’s relative weighing of price/quality attributes • Do not capture information from consumers past experiences • … No wonder they are not very successful (less than 5% of Internet users use shopbots) Opportunities for software agents • Personalized shopbots who adaptively infer individual consumers utility function – What factors matter most – Relative weighing of factors • Similar in spirit to recommendation agents Information Technology I Implications of “shopbots “shopbots”” • Benefits for customers – Better prices, service terms • Challenge for vendors – … but also helps vendors learn more about their competitors – Most vendors have responded with complex, rapidly changing price structures • Business opportunity for the mediating entity (the agent “operator”) – E.g. frictionless commerce. Com How to buy: Negotiation agents Information Technology I Intelligent Negotiating Agents BUYER Buy Agent Sell Agent User needs, Business sale criteria & & preferences pricing rules Negotiation about transaction SELLER Information Technology I How do agents negotiate? Information Technology I Implications of negotiating agents • Dynamic pricing becomes a reality – Everything is on auction • New, complex categories of auctions that were not practical before become possible – Combinatorial auctions where multiple “bundles” of goods are auctioned simultaneously – E.g. complete travel packages including airfares, hotels, tours, etc. Challenges • Standardizing the meaning of information • Trust building • Dispute resolution • Security • … Information Technology I When will all this happen??? • Historically, there has been a time-lag of about 15 years between the time that a major new technology has been proposed in the lab and the time it entered the mainstream of business • Agents were proposed in the late 80’s-early 90’s • Therefore, they are about to enter the mainstream by 2005!!! • (Historically, technological predictions involving time have been most unreliable)