Published: January 30, 2023

Google’s new artificial intelligence solution for retail stores was a hit at NRF

The tech giant is betting retailers want AI that can impact this year’s bottom line.

If you stopped by Google’s booth at NRF’s Big Show earlier this month, you probably had the same experience we did: a space so packed it was hard to get near the shelves! Retailers and reporters (?) were there to check out Google’s newly unveiled in-store tool: a camera-enabled, AI-powered shelf checker solution.

The retailer-focused AI pulls from Google’s giant database—billions of items—to identify products and shelf space, feeding retailers with real-time visibility into where restocks are needed and what shelves look like.

  • If you’re envisioning a retrofitted grocery store with thousands of cameras to make this work, think again. Google says implementation is a light lift, as retailers can use visual inputs from existing ceiling-mounted cameras (such as security cameras) from a mobile phone (operated by an associate) or from a store-roaming robot (...if you have one).
  • The shelf-checking AI uses two machine-learning models—a product recognizer, and a tag recognizer—to identify products and empty shelf space from “different angles and vantage points,” according to a release. So, yes, it can still identify that knocked-over box of Cheerios. And it knows you put that candy bar back in the wrong spot.


AI is something of a buzzword at the moment, with tools like ChatGPT and digital image generator DALL-E making headlines in just about every industry. But the managing director of retail industry solutions at Google, Amy Eschliman, says the shelf-checking AI has something the others don’t: immediate impact on the bottom line.

“Retail is, in many ways, at an inflection point with their use of AI,” Eschliman told Retail Brew at NRF. “We’re seeing retailers increasingly look towards, ‘How can I make a near-term impact with AI?’ How can I really either drive and improve customer experience, or think about my operations in a different way where I can make an impact this fiscal year?”

  • Eschliman pointed to February 2022 data from analytics company NielsenIQ, indicating that empty shelves cost retailers $82 billion in 2021, suggesting that Google’s new tool can help companies operate more efficiently.

Mind the store: Now, if an item isn’t on a shelf because it’s out of stock due to global supply-chain issues, camera vision and AI probably won’t help retailers refill their supply. But shelf visibility isn’t just a user-experience challenge; it’s also a key part of the frontline grocery worker experience, Eschliman said.

  • Retailers are thinking about how to not only retain workers by making their jobs easier, but about how to enable workers to be more productive, she said.
  • Traditionally, shelf-checking is a manual process, and even for stores that have automated it, the updates aren’t as frequent as they’d like, Eschliman added.
  • That’s where Google’s shelf-checker comes in: “It comes down to, ‘How can you help the associate understand what they should be working on? How can they work on the most important things that are going on in the store?’”—MA
© Public Gaming Research Institute. All rights reserved.