Shopping bots can be swayed by page tricks far more easily than people are — but not by the tricks you'd expect.
Do AI shopping agents change behavior based on how scenarios are framed?
This explores whether AI shopping agents make different choices depending on how products and tasks are presented to them (prices, ratings, nudges, page position, description wording, and how vaguely the shopper states their goal), and how that compares with human shoppers.
This explores whether AI shopping agents make different choices depending on how a shopping scenario is presented to them: the cues on the page, the order of the options, the wording of product descriptions, and how clearly the user explains what they want. The corpus says yes, and the effects are often stronger than they are for humans. In a benchmark that ran 17 language models and a group of people through the same binary product choices, the agents changed their choices far more than the humans did when prices, ratings or nudges were adjusted. Humans moved a little, in a probabilistic way. Agents tended to snap into rule-like patterns, so one small change could flip the outcome completely Do shopping agents respond more strongly to price and rating cues?.
The surprise is which framings work. Many of the usual retail persuasion tactics, such as scarcity warnings, countdown timers and bundle offers, had inconsistent effects across models. Star ratings and price were the only cues that reliably moved agents across thousands of simulated shopping rounds Do human shopping tactics work the same way on AI agents?. Wording still matters, though. When a seller-side AI rewrote product descriptions using competitor data, it raised the product's market share by roughly 4 to 15 percentage points with five of six buyer models. Price, rating and page position stayed the same, so the change came from the text alone Can AI seller agents shift buyer preferences through description edits?. Layout matters too. Agents favor certain list positions even in text-only interfaces and even when told to ignore position. Each model has its own pattern, and the pattern changes when the model is updated Do AI shopping agents show persistent position bias in text interfaces?.
The most useful finding for understanding why this happens: framing mostly matters when the agent can't see everything easily. Eight commercial models resisted marketing manipulation when all product details were visible at once. When looking up product attributes took extra effort and the user's goal was vague, the same models fell back on price-focused shortcuts. The breakdown happened in how the agents gathered information, not in their reasoning or arithmetic When do pricing cues actually mislead AI shopping agents?. In other words, you shape an agent's choices partly through what's on the page, and partly through how costly the page makes it to look past the first cue and how clearly the person states what they want.
This sensitivity raises questions about who controls the framing. If an agent's choices depend this much on presentation, then whoever writes the descriptions, orders the results or sells placements has a lot of influence. One former Visa data chief warns that agents funded by advertising will be pushed toward paid placements regardless of how well they're built Do advertising-funded shopping agents favor paid product placements?. Others argue that the influence runs in both directions. Through repeated conversations, the agent also shapes what the shopper ends up wanting, so framing affects the human's preferences as well as the agent's choice Do AI shopping agents reshape what consumers actually want?. This fits the argument that real market change requires agents on the seller side too Do one-sided consumer agents actually change market structure?. The description-rewriting experiment shows that seller-side agents already exist, and they work by reframing.
Sources 8 notes
A benchmark testing 17 LLM models against humans on binary product choices found agents exhibited much larger choice shifts in response to price, rating, and nudge interventions—collapsing into rule-like patterns rather than the modest probabilistic shifts humans showed.
A study simulating thousands of shopping rounds across four AI models found that classic behavioral tactics produced inconsistent effects. Only star ratings and pricing showed reliable influence on agent choices across models, while scarcity messages, timers, and bundle offers varied unpredictably.
A seller-side AI agent rewriting product descriptions based on competitor data increased focal product market share by 3.66 to 14.89 percentage points across five of six tested buyer models. The effect came from text edits alone, without price, rating, or position changes.
ACES found that AI agents exhibit strong, model-specific position biases even in headless text-only interfaces with explicit instructions to ignore position. The bias is a deep-seated model characteristic, not a superficial visual artifact, and recurs after each model update.
Eight commercial LLMs avoided marketing manipulation when all product details were visible, but fell into price-focused heuristics when attribute lookup was expensive and user goals were vague. The failure was in information retrieval, not reasoning or arithmetic.
Show all 8 sources
Bob Hedges, former Visa Chief Data Officer at MIT, warns that advertising business models create structural incentives for shopping agents to favor paid placements over neutral recommendations, independent of technical design quality.
Bilgihan and colleagues argue that transactional AI agents—which execute purchases rather than just recommend—end the consumer's sole cognitive role in preference formation. Preferences emerge through bidirectional feedback loops between agent learning and consumer trust, not independently beforehand.
Choudary argues that consumer-only agents like Muse and Instinct change where buyers enter markets but cannot alter market mechanisms themselves. True structural change requires seller-side agents too, enabling negotiation between represented parties rather than buyer optimization within fixed supplier offerings.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- Shopping by algorithm: How agentic AI deploys human heuristics as a surrogate consumer
- When Agents Shop for You: Role Coherence in AI-Mediated Markets
- A Framework for Studying AI Agent Behavior: Evidence from Consumer Choice Experiments
- What Is Your AI Agent Buying? Evaluation, Biases, Model Dependence, & Emerging Implications for Agentic E-Commerce
- Vertical tacit collusion in AI-mediated markets
- Are Behavioral Pricing Tactics Dying? A Study of Price Endings & Partitioned Pricing Effects on Machine Buyers
- Algorithmic Collusion by Large Language Models
- From Product Search to Preference Articulation: The Economics of Agentic Commerce