Meet the nominees for the 2022 VentureBeat Women in AI Awards!

July 8, 2022

Dave Woollard is Standard AI’s SVP of Technology Strategy and Innovation. A former NASA engineer and longtime computer vision expert, Dave is deeply involved with Standard AI’s strategic initiatives. 

Below is a transcript of a discussion about Dave’s perspective on the announcement that Standard AI is acquiring Skip.

Q: Connecting self-checkout with ceiling-mounted, autonomous retail platforms is an interesting concept. Can you explain more about it?

Retailers are in a difficult position today.

Everyone on the self-checkout side is basically trying to force this false dichotomy that it’s a question of self-checkout versus autonomous. And they point out that while retailers might not love self-checkout, it’s here today and can provide immediate labor relief. They say it’s the easy and practical answer. 

But when you put yourself in the shoes of a retailer, you’re not just making an investment in growing your tech stack or getting in bed deeper with an existing vendor. You’re also committing to a multi-year capex amortization. That’s because the reality is if you make a physical investment in the modification of your store, then you’re not going to be excited about doing that again in three years.

The acquisition of Skip, after closing, will allow retailers to do both self-checkout and autonomous checkout on a single platform. If you’re not yet convinced that autonomous is right for you—or you’re still trying to understand what it means for your company—then you can build a self-service experience into your stores, today, that has an obvious path to the future. You won’t have to replace hardware or redesign the layout of your stores if, or when, you decide to introduce autonomous technology.

I think it’s really powerful to be able to say that a single hardware investment today will yield dividends in the future. 

This is an opportunity to praise the Skip team, as well. One of the big selling points of their kiosks is that they don’t have a large footprint like the giant, legacy self-checkout machines that we’re all familiar with. They’re very sleek, modern, and easily-configurable. 

It’s very fitting with Standard AI’s ethos of “retrofit the store.” 

Q: Where will Standard AI get the most value from the acquisition? 

There’s an obvious headline here that we’re buying a self-checkout company. I certainly don’t want to lose sight of the important point that we’re meeting shoppers where they are and introducing autonomous retail through a familiar element of the customer experience. 

But for me, under the hood, one of the most important things that we’re doing here is the acquisition of a cloud-based point-of-sale (POS) offering—one that already has produced back office and loyalty integrations. That means we’re able to move much more quickly in the enterprise space. 

Retail technology stacks are notoriously difficult to integrate with, and legacy POS systems are not designed in a way that can be easily decoupled for modern technologies like ours. 

That may not be the headline, per se, but it’s where I see the most value.

Q: We know that this is a step forward for Standard AI, but some people might initially think this is a step backward for an autonomous retail company. What are your thoughts? 

We firmly believe that autonomous shopping is the best experience that can be offered to both customers and staff. With Vision OS^, we’re able to offer retailers a complete operating system for the store itself. 

I see self-checkout as a bridge technology that allows for shoppers to experience some level of convenience over the status-quo. It’s a good transitional technology to educate the shopper about autonomous shopping and how it works. 

Most convenience retailers’ portfolios have a heterogenous customer experience. Their stores vary not just in format, but also in the types of customers they serve at different locations. We can’t really believe that one solution will fit everyone across the board. 

By offering a wider overall feature set that allows us to address shoppers of all walks of life—and give retailers the ability to introduce autonomous shopping using their existing Skip self-checkout kiosks—I think this is a good opportunity to start bringing those shoppers on an autonomous journey. 

Q: Some companies are putting computer vision into the self-checkout hardware itself. Is that the future, or is the future computer vision on the ceilings of stores?

There are now a number of self-checkout solutions on the market that put computer vision directly into the hardware. Sort of like computer vision in a small box. The customer sets their items down, the device is supposed to identify them, and then the customer selects a form of payment.

I think this type of technology may ultimately prove to be a neutral value proposition for shoppers.

At best, there’s a novelty factor and the hardware delivers marginal improvements on speed that shoppers don’t really notice. The experience is still one where shoppers have to queue and wait their turn to pay. That’s the happy path, and it’s nearly identical to the status-quo.

The problem is retailers don’t consider the unhappy path, which is much worse than the status-quo. When those solutions don’t work—when the hardware struggles to identify items, when it’s confused, when it tells shoppers to separate overlapped items, or when shoppers have to call over store staff for help—that’s when they say “forget it” and do things the manual way at a traditional register. This, coupled with the fact that this hardware is more expensive than the status-quo, ultimately proves to be an expensive proposition for the retailer.

Convenience retail is elegant chaos. We’re looking for solutions that work—not novelty. And the future isn’t waiting in lines. Autonomous retail platforms assemble virtual carts on all shoppers, eliminating the need to do any of that. Grab what you need, and you’re good to go.

It’s worth pointing out that ceiling-mounted computer vision gives the retailer a breadth of insights and capabilities that you can’t get from computer vision in self-checkouts.

How do shoppers interact with those products prior to placing them on the bed of the scanner? If the cameras can’t see outside of the box and understand the full contextual journey throughout the store, they can’t tell you that. Nor can they tell you if someone spent a minute in front of the Doritos but ended up buying Frito Lays. 

For the investment in technology, I think it only makes sense to build a solution that has a future roadmap and growth in all areas of the store—not just the moment of checkout.