Week 12
Consumer Behavior · Davood Wadi, PhD
Consumers increasingly shop through screens, algorithms, and AI agents. The question is whose interests these intermediaries serve.
Most consumers now research, compare, and buy at least part of their purchases online. The smartphone is the main shopping device for many consumers. They check prices and reviews even while standing in a physical store.
Digital shopping leaves a record of every search, click, and purchase. Firms use this record to personalize products, prices, and messages.
Between the consumer and the product now stand several intermediaries: platforms, recommendation algorithms, and AI agents.
Mobile shopping is buying through a smartphone app or mobile website. Mobile shopping makes purchases possible at any moment, which increases impulse buying.
Social commerce is buying directly inside a social media platform, without leaving the app. A consumer sees a product in a video, taps it, and buys it in the same feed.
Live-stream shopping combines entertainment, a host's demonstration, and a limited-time offer.
Social commerce joins social influence and the purchase in a single moment, so the prepurchase and purchase stages almost merge.
Each purchase happens in a different way. Name the kind of shopping it shows.
Buying through a smartphone app, at any moment.
Buying directly inside a social media platform, in the same feed.
Entertainment, a host's demonstration and a limited-time offer together.
A mobile website makes a purchase possible at any moment, which invites impulse buying.
Social influence and the purchase meet in a single moment inside the platform.
A live demonstration with a limited-time offer.
Which concept does this case show?
A recommender system is an algorithm that suggests products based on a consumer's past behavior and the behavior of similar consumers.
"Customers who bought this also bought" and a personalized streaming home page are recommender systems.
Recommendations help consumers find relevant products in very large assortments.
Recommendations also narrow what consumers see. A consumer may never discover products outside the algorithm's predictions.
A filter bubble forms when the algorithm keeps showing more of what the consumer already likes.
In algorithm-driven discovery, the algorithm decides which products reach the consumer's attention.
Over a year, a streaming app's home page fills with crime documentaries after a viewer finishes several of them. The viewer never sees the app's comedies or nature series, although the app has thousands of them. What does this show?
Agentic commerce is shopping in which an AI agent acts for the consumer, not only advises.
At the first level, the AI recommends products, and the consumer searches and buys.
At the second level, the AI searches and compares options, and the consumer chooses and buys.
At the third level, the AI selects a product and prepares the purchase, and the consumer approves it.
At the fourth level, the AI completes the purchase on its own within limits the consumer has set, such as a budget.
At each level, the consumer hands more of the decision to the agent and sees less of the marketplace directly.
Marketers then inform two parties: the consumer who sets the goal and the agent that carries it out.
Each case shows how much of the purchase a consumer hands to an AI agent. Name the level.
The agent completes the purchase on its own, within a budget the consumer set.
The AI only recommends; the consumer searches and buys.
The AI searches and compares; the consumer chooses and buys.
The AI selects and prepares the purchase; the consumer approves it.
The consumer still approves before the purchase goes through.
Searching and comparing are handed over; choosing and buying are not.
Which concept does this case show?
In a principal–agent relationship, an agent acts on behalf of a principal and owes that principal a duty of loyalty.
A consumer who relies on an AI shopping assistant expects to be its principal.
The platform that deploys the assistant may also profit when the assistant recommends sponsored listings, which are products whose sellers paid for placement.
The AI agent then faces a conflict of duty between two principals: the consumer and the platform.
In hotel-booking experiments, an AI agent told that it worked for the traveler chose a sponsored hotel less often than an otherwise similar hotel without sponsorship (Wadi & Ma, 2026c).
When the same agent was told that it worked for the booking platform, it penalized sponsored hotels less (Wadi & Ma, 2026c).
Adding one sentence stating that its recommendations should serve the traveler removed this difference (Wadi & Ma, 2026c).
Whom the agent is told it serves shapes the recommendation the consumer receives.
A booking platform deploys an AI assistant whose instructions say only that it works for the platform. Travelers complain that it recommends sponsored hotels too often. Based on the hotel-booking experiments, which change is most likely to reduce this?
Disclosure rules require platforms to label paid placements, so that consumers can discount them.
When an AI agent reads the listings for the consumer, the label reaches the agent, and the consumer may never see it.
The wording of the label mattered. "Sponsored" reduced choice of the paid listing more than "Promoted" did (Wadi & Ma, 2026c).
A clearer label did not remove the difference between an agent that served the traveler and one that served the platform (Wadi & Ma, 2026c).
Disclosure rules written for human consumers cannot by themselves protect consumers in AI-mediated commerce.
Stating that the consumer is the beneficiary of the agent's recommendations offers a direct protection.
Firms collect data on consumers' searches, locations, purchases, and conversations with AI assistants. Personalization gives consumers relevant offers and saves them time.
The privacy paradox is the gap between what consumers say and what they do. Many consumers say they value privacy, yet they share personal data for small benefits.
Surveillance capitalism describes a business model that turns data on human behavior into predictions that are sold to others (Zuboff, 2019). Conversations with AI assistants can reveal needs, worries, and plans in more detail than a search history.
Ethical data practice asks consumers for informed consent, collects only what is needed, and explains how the data will be used.
Each situation shows one idea from the slide. Name it.
What the consumer says about privacy differs from what they do for a small benefit.
Data on behavior are turned into predictions that are sold to others (Zuboff, 2019).
A relevant offer built from the consumer's own data saves them time.
Informed consent, only the data needed, and an explanation of its use.
Stated privacy concern, then personal data shared for a small benefit.
The firm uses the record of the consumer's behavior to make the offer relevant.
Which concept does this case show?
Algorithmic pricing uses software to set and change prices automatically.
Dynamic pricing changes prices with demand and supply, such as surge pricing for rides or airfares that rise as seats sell.
Personalized pricing charges different consumers different prices for the same product, based on data about each consumer. Consumers often judge personalized pricing as unfair, especially when they learn that someone else paid less.
In simulations, pricing algorithms learned to keep prices high without any communication between the firms (Calvano, Calzolari, Denicolò, & Pastorello, 2020). When AI agents shop for consumers, pricing algorithms may face buying algorithms, and both the price and the choice may be set without a human in the moment of purchase.
Sort each price change: does it follow demand and supply, or the data about one consumer?
Tap a card, then tap the group it belongs to.
Sustainable consumption is buying and using products in ways that meet present needs without harming the ability of future generations to meet theirs.
The conscious consumer considers the environmental and social effects of a purchase, not only its price and performance.
The attitude-behavior gap is the difference between positive attitudes toward sustainability and actual purchases. Many consumers say they prefer green products but buy the cheaper option.
Greenwashing occurs when a firm makes misleading claims about the environmental benefits of its products. Resale, repair, rental, and sharing extend the life of products and reduce waste.
A consumer who delegates shopping to an AI agent can include sustainability in the instruction. The agent can then apply the criterion to every purchase, provided it gathers the information needed to judge it.
Each case shows one idea from the slide. Name it.
A misleading claim about the product's environmental benefits.
A positive attitude toward sustainability, but the purchase follows price.
Repair extends the product's life and reduces waste.
The consumer weighs the environmental and social effects of the purchase.
Rental lets many people use one product.
The firm implies environmental benefits the product does not have.
Which concept does this case show?
This course has followed the consumer from perception and memory through motivation, attitudes, decisions, and social influence.
The same theories now describe two actors: the consumer and the AI agent that searches and buys on the consumer's behalf.
The agent perceives products through data, retrieves brands from what has been written about them, and can show heuristic-like choices when information is costly.
The consumer still holds the goal, the values, and the experience of using the product.
Marketers who understand both actors can inform them honestly. Marketing that exploits the limits of either actor harms consumer well-being and weakens trust in the market.
Imagine you give an AI agent the authority to do your weekly shopping for one year. Which goals, limits, and values would you write into its instructions? Whose interests should it serve if the platform that runs it earns money from sponsored products?