Instacart ends AI pricing tests after FTC probe

Instacart has stopped allowing retailers to run AI-driven pricing tests on its grocery delivery platform after facing scrutiny from regulators and lawmakers.

The company acquired Eversight, a provider of pricing experiment software, for US$59 million in 2022.

A recent study by Consumer Reports and others found that Instacart’s algorithmic pricing led to shoppers paying different prices for the same items at the same store, with cost variations of about 7% per basket.

The Federal Trade Commission has sent a civil investigative demand to Instacart regarding its pricing practices.

Instacart said retailers, not the company, set prices on its app, and denied using personal or demographic data for the pricing tests.

🔗 Source: CNBC

🧠 Food for thought

Implications, context, and why it matters.

Instacart’s AI pricing tests were limited in scope, not across the whole app

Consumer Reports’ tests with 437 volunteers at Safeway plus Target saw price swings up to 23% on the same items 1. Instacart said the trials were limited and short-term, with randomization 2.
Instacart bought Eversight in 2022 and folded it into ads plus retail tools 3. In 2021 it set records for orders and gross transaction volume 3. Revenue, ad revenue, and gross profit also hit highs 3. That backdrop supported a bigger pricing and promotions push.
Retailers used the tests to try small price changes that lifted margins by up to 6% in measured categories 3, but only 10 partners joined 1, so the gross merchandise value (GMV) effect likely stayed small versus Instacart’s total volume.

Compliance tools gain ground as FTC scrutiny pushes retailers toward verifiable pricing systems

The Federal Trade Commission (FTC) sent orders to eight companies on surveillance pricing 4 and warned that using algorithms to dodge price-fixing laws is illegal 5, which pushes demand for tools that verify pricing code does not collude or use banned personal data.
B2B vendors can ship pricing audit software with human override, anonymized training sets, and proof of procompetitive effects 5 that keeps dynamic pricing while lifting margins by 2-5% 1.
Law plus compliance advisors can win work with risk reviews of current pricing tools, as courts now accept antitrust claims when AI uses shared nonpublic rival data 6, and European regulators plan to widen cartel rules to cover AI-driven collusion 7.