AiFi’s demo of its Autonomous Store Platform at Mobile World Congress LA. 

Imagine walking into a grocery store, picking up all the items you need, and walking right out the door. No need for a cashier or even a self-checkout, the store will automatically charge everything you take to your credit card.

AiFi, an AI technology company based out of Santa Clara, Calif., has launched a system – the Autonomous Store Platform – that it says can fully automate stores of any size from pop-ups and kiosks all the way to big-name retail stores. Combining an array of off-the-shelf RGB cameras with the company’s own computer vision system it calls Open Autonomous Store Infrastructure and Services (OASIS), allows stores to track customers and the items they pick up – giving consumers a streamlined, automated shopping experience while also letting stores track inventory and sales analytics.

The company was on hand at Mobile World Congress (MWC) 2019 in Los Angeles giving live demos of its technology and a look at its backend operating system. Trade show demos always represent pretty ideal conditions, but the the shopping experience was very seamless. After scanning into the store with an app for entry (the app would contain your payment information), you simply proceed through the store and any item you pick up is automatically identified by the OS and added to your shopping cart. There’s no need to scan a barcode of any item and it doesn’t matter how you pick it up. Upon leaving the store anything you carry out with you is charged to your account.

Speaking with Design News at MWC LA, Liu Yang, a representative from AiFi, said the store is able to track items so easily because AiFi’s computer vision system has been trained on recognizing every item in a store’s inventory in various orientations and lighting conditions. “We’re looking at items the same way human eyes are looking at items,” Yang said. “We actually build out the system to streamline the process as well. For example, if the packaging of an item changes or you’re introducing a new item into the system a store manager can just use a phone to take pictures of the item and upload them via our tool.” The clear advantage here is a store doesn’t have to take down its entire operation to have its algorithm retrained to accommodate an inventory change.

As far as tracking individual customers Yang said AiFi is sensitive to privacy protection concerns and that their system does not track people based on ID or personal appearance. AiFi does not employ facial recognition technology. “We recognize you by the key points on your body – your joints, legs, ect.,” Yang said. The system tags each shopper with a pseudo ID as they enter a store and deletes it when they leave.

“Right now we run our system on the edge, so no data leaves the store,” Yang said. “This is for privacy protection and to comply with regulations like GDPR in Europe. Also, most retail stores do not have the infrastructure to handle the high speed Internet for the system to be cloud-based.” AiFi is counting on the emergence of 5G and increasingly more powerful edge computing to fully facilitate systems like OASIS in the coming years.

A visualization of AiFi’s OASIS as it tracks shoppers through a store. (Image source: AiFi)

Just imaging the idea of autonomous shopping should have hackers drooling thinking of clever workarounds to trick the system. Certainly anyone with access to your phone and your login information on a store’s app could get up to some dirty business, but that’s nothing new from today’s retail experience.

And while no system is completely hack-proof AiFi remains confident that OASIS is accurate enough to keep track of individual shoppers and their items. Yang added the company is also not looking to fully automate stores and that humans would still be used for inventory, restocking, security, and management of stores. “The whole thing is just way more efficient. A team of four people could manage multiple stores simultaneously. And our system will send alerts if anything is out of stock or other things need attention.”

AiFi has already launched pilot programs with various grocery and convenience store chains around the world including Carrefour supermarkets in France, Albert Heijn in the Netherlands, and Żabka in Poland. The company is also collaborating with Valora, a convenience store chain in Switzerland to open a series of “nanostores” – tiny, fully-automated shops akin to what you might find in a hotel lobby – that operate 27/4.

A video demonstrates AiFi’s nanostore concept – small stores that are fully automated by AI and can run 24/7. (Source: AiFi)

Yang said the company is also conducting ongoing work to test the robustness of OASIS. It may be able to handle a typical level of store traffic, but what happens during a big Labor Day Sale or, even worse, Black Friday?

The company is also exploring other, more crowded, use cases. “We’re looking at scenarios like airports and train stations as well,” Yang said. “The store we just launched with Albert Heijn in the Netherlands is going to be in airports and public transit areas. We already have test stores open to Albert Heijn employees so the next step is opening them to a larger audience.”

AiFi declined to share any more information about its pilot programs. Yang said European unions have expressed concern about automation’s impact on the workforce so the company has been cautious about making announcements.

Services like Instacart and DoorDash and various delivery options offered by Amazon have already eaten away at the retail sector in terms of getting actual humans in stores. In this context AiFi’s system looks like a big bet into a future where consumers are still willing to visit brick and mortar stores.

Yang said AiFi is envisioning a world where stores have become more localized to specific neighborhoods and towns to offer consumers more convenience – hence the nanostore concept. “Stores are getting closer to us with more relevant inventory for all of us,” Yang said. “There’s still going to be scenarios [for stores] though. For instance, imagine you’ve just taken a red eye and you arrive at the airport 4AM. Do you want something warm or are you going to wait for a package? Or think about remote regions as well or gas station stores…We envision in the future shopping will be a 24/7 convenience and more personalized.”

Chris Wiltz is a Senior Editor at  Design News covering emerging technologies including AI, VR/AR, blockchain, and robotics.