AI on the move
RetailNow panel calls tech players to action
Anyone active in the retail space knows that AI, in addition to improving the customer experience, is changing many if not most retail functions. In order for retailers to adapt to this new reality, they need the support of technology providers and a game plan for introducing AI tools.
“As we all know, AI is no longer simply a buzzword,” Kathy Crumley, global strategist at Intel Corp., said during last week’s RetailNow show in Las Vegas at Caesars Palace. “It is something that we’re all experiencing. It’s coming in fast and furious, and a multitude of different capabilities.”
Kathy Crumley, Intel Corp.
Crumley explored the opportunities AI can deliver and the challenges retailers face in addressing them during an educational session, “The Omnichannel Opportunity; AI, Customer Experience and the Human Touch.” The session featured six technology panelists.
Panel cites use cases
She began the session by asking each panelist to give a use case where AI delivers immediate value, results and ROI. Their answers demonstrated the wide range of AI benefits at retail.
“Inventory visibility is key in retail right now,” said panelist Dean Reverman, vice president of marketing at BlueStar, a distributor of B2B electronic devices that sells exclusively to value added resellers. “You have to be able to analyze your inventory, understand what your inventory is doing, what’s coming in, what’s going out.”
Panelist Russell Harty, North America president at LANDI Global, a payment technology solutions provider, cited the smart mirror. A virtual smart mirror can act as a fashion consultant that recommends apparel for an item a customer is trying on in the store. Harty said the device is especially helpful for increasing upsells at the point of sale.
“You go into an apparel retailer, you’re trying on a shirt or whatever,” he said. The smart mirror immediately recommends other items, such as shoes, that match what they’re wearing. “For the ROIs, part of it is the upsell.”
Panelist Duffy Fron, executive director of sales at Toshiba Global Commerce Solutions, cited loss prevention, thanks to customer purpose recognition at the point of sale. The MxP Vision kiosk designed with Toshiba’s Elera IoT security suite reduces shrink and friction at self-checkout. Integrated with smart pad and edge camera technology, the kiosk provides the ability to monitor customer behavior and prevent loss activities such as double stacking.
Fron said Toshiba was able to help one retailer speed up the purchase process by 30 seconds and save $2 million a year in produce loss.
“Most nefarious actors start their process by looking this way and that way, and maybe looking up…and it (the technology) was able to see that,” Fron said. “We were able to reduce shrink and we were able to speed up processes.”
Panelist Mike DeMarco, vice president of sales at meldCX, an enterprise platform dedicated to digital transformation, also cited the reduction of in-store theft. He said AI powered cameras in self-checkout kiosks alert customers to missed scans and alert an employee near the kiosk.
Panelist Zahdan ELZahdan, vice president, North America sales and global accounts, Pyramid Computer, cited the improved ability to track items on demand and better identify customer buying habits.
“Technology allows you to monitor not only what’s happening in the facility, but also impact what can happen in your facility with the environment outside,” ELZahdan said. “If there’s a blizzard coming, a retailer selling snow shovels will be able to measure what impact the blizzard has on sales of that item.”
Retail functions impacted
Crumley presented a slide listing store functions where AI can be used:
- POS
- Produce recognition
- Self checkout
- Assisted checkout
- Alternative checkout (curbside pickup, buy-online-pick-up-in-store, home delivery)
- Loss prevention
- Line busting
- Smart cart
- Deli ordering
- Consumer mobility
- Loyalty and promotions
These functions address the use cases available to the consumer and the considerations the consumer is going through, Crumley said.
“They’re obviously going to pick the path that is correct and right for them, but at the end of the day, mapping this out and then taking down and really getting on the use cases for each one of these steps is something that we really glom deep into and look into,” she said. “It’s all being driven by decisions that are happening in the shopper experience.”
Crumley said it’s important to tie the data metrics to the consumer transactional experience.
“If you can’t measure it, you can’t manage it,” DeMarco agreed.
AI retail game plan
Asked how businesses should align AI with business goals, ELZahdan said a business should not invest in AI for the sake of doing so. AI should add measurable value to some aspect(s) of the operation, such as improving personalization or streamlining operations.
“Too often, AI gets labeled as an innovation project, really disconnected from actual business outcomes,” DeMarco said. Instead, a retailer should consider what would make a customer’s decisions faster, better or cheaper. Once it is deployed in operations, the business should be able to measure the results.
MeldCX is currently working with 50 Starbucks stores with AI powered cameras to calculate customer dwell time and other metrics, such as tracking how many large cups were sold.
“All these different insights (are) from a simple AI camera above the store,” DeMarco said, adding that the cameras do not store any customer privacy data.
Retail education imperative
The panelists agreed that education about AI is needed in the retailer community. Here is an area where technology providers need to step up to the plate and address objections.
“Pick a lane, and develop an ecosystem,” Reverman said. “You have to have partnerships in order to be successful.”
“I would advocate that you at least get off the block,” he said. “You can’t remain stagnant and think that if you don’t get into the AI game, all is going to be well. You have to get off the block.”
Technology providers can enable retailers to make better use of the data they are already collecting, Crumley said.
“The data is being captured, it’s just not being utilized,” she said. “The retailers have the data…The queries that need to start to happen at the enterprise retailer perspective is to tease out all of those factors that are really going to drive the SL (store leader) opportunities.”
Crumley presented a slide showing that retailers that use AI are growing sales 2.3 times as fast as those that do not, combined annual growth rate of the AI retail market is 23.9%, and AI spend at retail is expected to jump 400% by 2026.
The upsell opportunity, personalization
ELZahdan expanded on the importance of the upsell opportunity. AI can access information about the consumer’s browsing habits that enables the upsell opportunity significantly, because it enables the retailer to reach the consumer at the point of sale.
To this point, Crumley said Kroger is doing a good job in this area. “They will map out what I’m buying, and send me promotional items in the mail,” she said. “I think they’re doing it good. There’s always room for improvement.”
Reverman expanded on the benefits of personalization. “There’s a lot of technology that’s enabling that,” he said. “Making it so that when I go to pick up my order, it’s not just the receipt that’s saved on there but it’s personalized to me and that whole experience.”
Computing considerations
The panelists agreed on the importance of edge computing for deploying AI.
“You need to have a compute at locations to enable these low latency type of AI applications that are riding there on the edge,” ELZahdan said. “Gone are the days when you have live towers in the back room, one in the POS and one in the security cameras…all that is being compacted by the single compute that can then run other AI agents.”
Crumley agreed. “Edge computing is one topic that from an enterprise retail perspective that doesn’t get enough promotion, because I think it’s been a scary topic for a lot,” she said. “Gone are those days of having a plethora of servers in the back room. You may have one and it’s got enough power and wiggle room and it’s going to let you grow into the future, but it’s a consolidated workload type of server and device…From a camera to checkout to a point of sale or to a kiosk or a digital sign.”
“And we’re not saying you have to abandon the cloud based initiative,” Crumley continued. “The two coexist. What we’re saying is there has to be accommodations and considerations for edge computing simply because of the amount of data. There’s more data and you need the infrastructure to do that.”
Partner considerations
In the process of educating retailers about AI, it’s important for retailers and technology providers alike to carefully evaluate what technology partners to work with.
“Does the company you are transacting with match your DNA?” Reverman asked, adding that the partner must understand the channel you work in to provide the support you need.
The panelists agreed that the process of educating retailers about AI and working with them to introduce it is not an easy challenge, and mistakes should be expected along the way.
Cover photo: Seated at left are: Dean Reverman, BlueStar; Russell Harty, LANDI Global; Duffy Fron, Toshiba Commerce Solutions; Zahdan ELZahdan, Pyramid Computer; and Mike DeMarco, meldCX.
