According to Gizmodo, the Federal Trade Commission has opened a probe into Instacart’s pricing practices after a report found the company charged some customers nearly 25% more for identical groceries. The investigation, confirmed by Reuters, focuses on Instacart’s “Eversight” tool, an AI-powered pricing platform it acquired in 2022. The scrutiny stems from an October 2024 experiment by Groundwork Collaborative, Consumer Reports, and More Perfect Union, where volunteers found price variations of up to 23% on the same items from the same store at the same time, with some carts paying an average of 7% more at checkout. Instacart claims these were randomized tests run with a small group of retailers and were not based on customer demographics. Following the report, Instacart published a company update titled “Instacart’s Commitment to Affordability,” promoting its work on price parity between online and in-store shopping.
The AI tool at the center
Here’s the thing: the language Instacart used to sell this Eversight tool to retailers is pretty damning. On its website, until at least October of this year, it marketed the software as a way to “optimize your pricing with AI” and to “continuously drive growth with dynamic pricing through experimentation.” They told retailers it could boost sales by 1% to 3% and margins by 2% to 5%. Now, Instacart is in full damage-control mode, telling Gizmodo this is “not dynamic pricing” and that prices don’t change in real-time. But come on. When you sell a tool on its ability to experiment with pricing to drive growth, what did they think was going to happen? The archived marketing page says it all. It’s a classic case of the tech being sold one way to business partners and explained another way to the public and regulators when the spotlight gets too hot.
The big picture problem
So, is this illegal? That’s what the FTC is trying to figure out. Instacart’s defense hinges on the randomness of the tests. They say it’s not based on your personal data, so it’s not discriminatory in a traditional sense. But that’s a pretty low bar, isn’t it? The core issue is transparency, or the utter lack of it. Customers have no idea they’re being used as guinea pigs in a pricing experiment. They just see a total at checkout. In an era where grocery inflation is a massive pain point, using AI to secretly test how much more you can squeeze out of people feels particularly slimy. It erodes any semblance of trust. And let’s be real: if the tests are truly random and not optimizing for profit, why run them at all? The whole premise of the tool is to increase retailer margins.
What happens next
This FTC probe is a big deal. It’s not just about Instacart refunding a few bucks. It’s about setting a precedent for how AI can be used in pricing, especially for essential goods like food. The agency’s statement that it is “disturbed by what we have read” is a strong signal. Instacart’s quick pivot to talking about “price parity” is a clear attempt to change the narrative. But the cat’s out of the bag. The investigation will likely dig into how the Eversight algorithm actually works, what data it uses, and how “random” these tests really are. For an industry that relies on consumer convenience, getting a reputation for algorithmic price-gouging is a terrible long-term business strategy. Basically, they got caught with their hand in the cookie jar, and now they have to explain to the federal government why the jar was there in the first place.
