It isn’t easy for brick-and-mortar retailers to compete with internet behemoths such as Amazon, Jet, and eBay. Online shoppers spent a record $ 108.2 billion last holiday season, according to Adobe Analytics, and that number is only expected to grow. By 2021, Statista projects that more than 230 million customers will browse products, compare prices, or buy merchandise online at least once.

So how’s a big box store to compete? Jamie Rapperport, cofounder and CEO of Eversight, thinks the solution might be artificial intelligence-powered dynamic pricing. “We want to bring pricing experimentation and promos right to the shelf,” Rapperport told VentureBeat in a phone interview.

Enter Eversight Pricing Suite, a new cloud software solution that’s meant to complement Eversight’s core platform, Eversight Cloud, which allows brands and retailers to test, adapt, target, and validate in-store promotions before they’re rolled out at retail.

Here’s how it works: When a retailer chain gives Eversight Pricing Suite the green light, the platform automatically begins running experiments on a subset of items throughout a handful of stores, taking into account factors like competitors’ pricing and shoppers’ historical behavior.

The tests are randomized to ensure they aren’t impacted by external factors, and they’re run across a number of product categories. “It’s tremendously complicated,” Rapperport said. “[The AI] figures out the inner relationships of different items and looks at the impact of prices in a highly structured way.”

After at least a week has passed, the system evaluates the results, taking stock of the price’s impact on sales. If a group of stores sold more of an item after a price increase or decrease, and the difference is statistically significant, it’ll suggest making the new price point permanent. And if an item sold poorly, the system will provide insight into which product categories might be hurting sales.

Rapperport is adamant that Eversight Pricing Suite doesn’t discriminate against demographics of customers or specific markets. Prices set by the AI are published to all customers in a store, and prices are never increased above MSRP, meaning the system only tests discounts of different depths.

“We’re not treating people differently,” Rapperport said. “We’re [simply] giving retailers the tools to compete [with online stores] … with a different merchandising technique. We see it as a continuous learning machine that gets smarter on price everyday and gives retailers the power to beat their brick-and-mortar competitors, as well as pure-play ecommerce retailers.”

Eversight, which is based in Palo Alto, California, has raised $ 24.2 million to date in funding rounds led by Sutter Hill Ventures, Emergent Capital, and others. It counts Coca-Cola, Hershey’s, Mars, and Molson Coors among its clients.