This Backgrounder examines how AI is advancing algorithmic pricing and how consumers shop, the emerging policy response, and the implications for CEOs. Businesses have long charged different prices for the same product or service. Hotels and flights typically cost more during vacation seasons and holidays. Deck furniture is more expensive leading up to summer, and prices rise for snow gear as temperatures fall. However, modern computing and the popularization of e-commerce in the 1990s introduced new opportunities for retailers to use customer data not only to target marketing but also to offer different prices to different customers. In 2000, for example, Amazon shoppers discovered they were shown different prices for the same DVDs, which the company said was part of a randomized pricing experiment.1 Over time, pricing models became more sophisticated. In 2012, the Wall Street Journal found that Staples.com displayed different prices depending partly on a customer’s estimated location and proximity to competing OfficeMax or Office Depot stores.2 And in 2015, ProPublica found that the test prep company Princeton Review charged different prices for an online SAT course depending on the student’s ZIP code.3 Pricing In the AI Era However, the growing availability of consumer data and, more recently, advances in AI have accelerated the use of “algorithmic pricing,” which broadly refers to the use of automated systems to set, recommend, or adjust prices. Some systems rely on relatively simple rules – such as raising prices when inventory falls below certain levels – while others consider large volumes of data on competitor prices, customer behavior or characteristics, and transaction histories to set prices, often updating in real time. Use of these tools is already common. A 2024 survey of pricing managers found that nearly 68 percent worked for companies using pricing algorithms for at least some products.4 Among firms using them, the most common inputs were cost data and historical revenue or profit, while more than half used information about past consumer behavior that could support customer-specific pricing. AI-Powered Shopping and Agentic Commerce AI is also shaping the way consumers shop as more people use AI tools to generate product recommendations, price compare, and monitor products for discounts. A 2026 Visa survey, for example, found that 54 percent of US adults had used AI for at least one shopping task, most commonly for price comparison.5 In principle, such tools could introduce more price competition and benefit consumers. However, they also raise questions about whose interests those agents serve. One study of approximately 325,000 AI-agent interactions found that agents instructed to find the cheapest option sometimes used information about a consumer’s financial circumstances to instead recommend more expensive options.6 Separate testing of 13 large language models found that systems asked to identify the cheapest product sometimes returned sponsored products that were more expensive.7 The growing use of AI tools to not only recommend products but make purchasing decisions also raises new questions about conflicts of interest and accountability when they recommend or select higher-priced options. Agentic commerce could therefore create a new dynamic in which both sides of a transaction use AI. Sellers may deploy algorithms to determine the price and consumers deploy agents to compare prices, wait for discounts, or negotiate better terms. These trends are further complicated by decisions online retailers face over whether to allow third-party AI agents to access their platforms or develop shopping agents of their own. Amazon has blocked external shopping agents, including Meta’s Muse, while promoting its own shopping tools; eBay prohibits unauthorized “buy-for-me” agents and other automated shopping tools from placing orders without human review, although it allows approved partnerships. Other retailers, including Walmart and Shopify, have taken a more open approach to third-party agents. These differing strategies reflect a broader tension between gaining access to consumers who increasingly shop through AI and maintaining control over customer relationships, advertising, data, and the checkout experience. Policymakers are increasingly examining whether existing consumer protection and competition laws are sufficient to address personalized and algorithmic pricing, or whether new rules are needed. To date, the response has been fragmented, with Federal agencies and Congress emphasizing transparency and oversight while several states have moved toward more specific disclosure requirements and restrictions. Federal Level The Federal Trade Commission (FTC) has increasingly focused on personalized pricing through its consumer protection authority. In August 2026, the FTC invited public comment on a proposed enforcement policy statement describing how the undisclosed use of personal information to set a price unique to a consumer could violate the FTC Act.8 And in October 2026, the FTC announced an inquiry into nine firms requesting detailed information that will allow the agency to study the use of personal data to set prices and compensation. Congress has also increasingly focused on personalized and algorithmic pricing. In March 2026, the House Oversight Committee opened an investigation into the use of AI and consumer data to set prices and requested information from several travel and retail firms. The Senate Judiciary Committee also held a hearing in August examining how consumer data may be used to determine prices. Lawmakers have also introduced legislation ranging from disclosure requirements to sector-specific (e.g., residential rents and food delivery) prohibitions on the use of personal data to set prices. While these will expire at the end of the current Congress, similar bills are likely to be introduced next year. The use of personal data to set prices can also raise discrimination concerns when seemingly neutral characteristics correlate with protected classes. Variables such as ZIP code, purchasing history, education, language, or browsing behavior may serve as proxies for characteristics such as race, sex, age, or national origin even when those characteristics are not explicitly included in a pricing model. The legal implications depend on the product and statute involved. In consumer credit, for example, the Equal Credit Opportunity Act prohibits creditors from charging different prices or fees on the basis of protected characteristics, and those requirements apply regardless of whether the decision is made by a person or an algorithm. More broadly, businesses may also face exposure under other Federal or state civil rights and anti-discrimination laws where personalized pricing results in unlawful differential treatment. State Level The use of personal data to set prices has also drawn attention from state policymakers. New York’s Algorithmic Pricing Disclosure Act, which took effect in November 2025 and is currently being litigated, requires most businesses using personal data to set individualized prices to clearly disclose that the price was set by an algorithm using the consumer’s personal data. New York lawmakers have since passed broader legislation that would prohibit surveillance pricing while preserving certain discounts and other exceptions. The measure has passed both chambers but has not become law. Likewise, Maryland’s Protection From Predatory Pricing Act, which took effect in October 2026, restricts food retailers and third-party delivery services from engaging in specified dynamic pricing practices or using consumer personal information to set prices. And in July, New Jersey enacted a law restricting the use of personal information to set individualized prices for groceries and other food products while preserving traditional loyalty programs and other bona fide discounts. It also imposes a temporary moratorium on new uses of electronic shelf labels while the state studies the technology. Connecticut has adopted a broader law requiring online disclosures and restricting surveillance pricing by retailers and third-party delivery services while preserving pricing differences attributable to supply and demand, differences in the cost of serving customers, loyalty programs, broadly available promotions, and other specified discounts. Connecticut law also prohibits residential rental property owners and managers from using revenue management devices that calculate nonpublic competitor data to set rental rates or occupancy levels. The policy debate remains unsettled, but companies using automated pricing systems should prepare for greater scrutiny of both the information that goes into pricing decisions and the differences those systems produce. Understand How Prices Are Set Executives should understand whether their companies use algorithms to set or recommend prices, know what information those systems use, and be prepared to explain why prices may differ. Reviews should distinguish among inputs based on costs, supply, demand, inventory, and competitors’ publicly available prices and inputs derived from individual consumers. This may be more difficult than it appears. Pricing systems can incorporate hundreds of variables, while data used for marketing, loyalty programs, or customer analytics may also feed into pricing or promotion systems. Companies should maintain an inventory of material pricing systems and document which categories of information influence their recommendations. Evaluate Pricing Vendors Companies should evaluate third parties that provide pricing, revenue-management, customer-segmentation, or optimization technologies. Reviews should determine what data vendors receive, what outside data they acquire, how models combine customer and market information, whether systems can generate individualized prices, and whether companies can explain how recommendations are produced. Contracts should also address compliance responsibilities, data usage, audit rights, model changes, record retention, and the company’s ability to limit or disable particular pricing features. Prepare for a State Policy Patchwork National firms should prepare for different state approaches to disclosure, prohibited data, discounts, loyalty programs, and dynamic pricing. Businesses will need to determine whether to operate different pricing systems across jurisdictions or adopt common practices that satisfy the most restrictive requirements. The latter may reduce compliance complexity but could also limit pricing strategies in states where broader practices remain legal. Consider Consumer Trust Legal compliance does not eliminate reputational risk. Consumers may react differently to a price changing because demand increased than to learning that the same company charged another customer less because of differences in browsing history, location, or other personal information. Companies should evaluate how a pricing practice would appear if publicly disclosed and whether they can explain the rationale in straightforward terms. Engage Policymakers Companies can work to help policymakers distinguish harmful uses of personal data from ordinary dynamic pricing, targeted discounts, loyalty programs, experimentation, and algorithms that incorporate competitors’ data. They can offer suggestions for technology-neutral rules and also provide evidence about how disclosure requirements, restrictions on particular data, and differing state standards would affect prices, competition, innovation, and compliance costs.Trusted Insights for What’s Ahead®
From Dynamic to Personalized Pricing
Policy Landscape
Implications for CEOs
Endnotes