Convenience stores are integrating more artificial intelligence (AI) into daily operations, including formats such as audio and video. Retailers invest in these types of AI to improve labor, continuously train employees, stay on top of maintenance, push customer loyalty and more.
Alan Meyer, CEO of convenience-store chain Mach 1 Stores, takes pride in being a progressive retailer who only expands the business when it makes sense.
“I always question the people that judge that [AI is] not for them without trying it,” he said. “To me, software as a service is what is going to allow us to compete as these consolidators continue to grow and build their own software and allow scale to swallow us whole.”
He also points out that AI is only going to grow.
“You're going to want to get acclimated to it now, because eventually this is going to be the elephant in the room. If you want to turn a blind eye to it and said, ‘That's not for me,’ your competitors are going to use it. There's no doubt in my mind about that. I think it's going to be an absolutely critical part of operational execution.”
Audio AI
The convenience-store chain, based in Effingham, Illinois, has 23 locations throughout central and southern Illinois and Indiana, and it uses InStore.ai to monitor customer-employee interactions, maintenance issues and out of stocks at six of its stores.
InStore.ai is a software developer based in Los Altos, California, that gathers audio from microphones strategically placed inside stores and transforms it into actionable data.
The microphones are usually near the cash register, self-checkout, fresh food or drive-thru. The AI identifies and assesses customer and cashier interactions, such as a customer complaining about dirty bathrooms, something not functioning or a cashier’s attempt to increase the basket size or offer a promotion.
Meyer likes the idea of AI providing actionable plans based on the data it collects.
InStore.ai offers recommendations in a variety of different formats, such as a dashboard, emails with summaries and real-time alerts.
“Management acts a lot faster on things when they hear the true voice of the customer saying that they're frustrated that something's not working, and it turns out they get it fixed much faster,” said Jay Blazensky, founder and CEO of InStore.ai.
While there is a subset of customers who don’t speak their frustrations and some that voice their opinions more often than the average person, the sheer number of transactions at a convenience store per day cancels out those anomalies, Blazensky said.
Convenience stores serve 160 million customers per day, according to NACS data. “That's a lot of conversations every day between cashiers and customers,” Blazensky said.
InStore.ai can also deliver data to fuel brands that care about the customer experience in the stores that their brand is associated with. Consumer packaged goods (CPG) brands also have an interest in what customers are saying about their product category.
Video AI
Some convenience-store retailers have also invested in Standard AI, a technology company based in San Francisco. The company’s product suite, called the Vision Platform, provides retailers with analytics—such as low- or out-of-stock items, theft prevention and employee compliance—based on strategically placed AI cameras.
The cameras continuously run video against images from the store that show how the shelves are meant to be stocked. Empty facings trigger an analysis of whether the space needs to be restocked or if there is a compromised planogram. Both scenarios are good to know—whether it results in a real-time intervention or an analytical understanding that needs more review, said David Woollard, chief technology officer at Standard AI.
Standard AI takes a consultative approach when working with retailers, said Woollard. The team is balanced with technical experts and experts from the CPG, logistics and retail side.
There are also retailers and CPG brands who are interested in understanding conversion funnel, which involves a broader set of analytics. Standard AI works closely with embedded data science teams and business folks to test hypotheses about merchandising and packaging.
Maintenance and Merchandising
Like Meyer from Mach 1, Babir Sultan, president and CEO of four-unit, Independence, Missouri-based Fav Trip, was drawn to InStore.ai because of its ongoing maintenance alerts.
“We want to make sure we're meeting those needs,” Sultan said. “When you're running one store, it’s different than when you’re running multiple stores. It becomes an issue where the upkeep becomes difficult.”
Sometimes, cashiers opt not to tell management about maintenance issues, especially if they are already intimidated by them, said Blazensky, but InStore.ai notifies management automatically.
In addition to maintenance, merchandising is a continuous task for category managers. With Standard AI’s predictive shopper behavior, retailers can run simulations to see what would happen if they moved a product to a different portion of the store, Woollard said.
“[It allows retailers] to play with a digital model, a digital twin of their store from a merchandising perspective, rather than having to go through all of the logistical hurdles and reconfiguring a store and then experimenting and measuring,” he said. “We're able to do that in simulation with great fidelity.”
Standard AI offers a fresh foods tracker to gather real-time data and intervention opportunities for convenience-store retailers to stay on top of out-of-stock and spoiled food. It alerts employees about out of stocks in the fresh food, roller grill and warming drawer areas with an employee app.
It can also be used to collect data for the company to understand ongoing trends, he said.
For example, Woollard said, there was an instance when a store was supposed to have fresh food on the shelf by 5 a.m., but it wasn’t prepped and shelved until an hour later.
“We were able to quantify the loss associated with the missed opportunities of shoppers pausing in the fresh foods and realizing there was nothing available,” he said. “That’s the dual-sided aspect of that platform—both in the real-time intervention but also in the analysis to understand more strategically how to deal with operational changes.”
“To me, software as a service is what is going to allow us to compete.” —Alan Meyer, Mach 1
Loyalty Push
Mach 1 didn’t have an effective way to test how often employees pushed customers for loyalty enrollment until InStore.ai, Meyer said.
Sultan is also eager to use AI to capture how often employees are communicating with customers about joining its 13,000-member text club.
Sultan has found that the verbiage surrounding rewards makes a difference. For example, Fav Trip’s text club includes discounts, so instead of asking a customer if they are part of the text club, he encourages employees to ask if they would like a discount.
“That ends up creating curiosity, where [a customer might say] ‘Yes, I would. How do I sign up?’ Or, ‘How do you do that?’ That's very different,” Sultan said.
Compliance
The tobacco tracker is one of the c-store specific functions in Standard AI’s product suite, said Woollard.
“Tobacco involves a high-value asset and also is regulated,” he said. “There's a lot of compliance aspects, and there's also a lot of stakeholders that are active in management of those compliance issues. We can monitor store employee interactions with tobacco products, make sure those tobacco products end up in transactions and everything was compliant … It’s a win-win from a compliance perspective and retailers’ operation perspective.”
When it comes to customer theft, real-time intervention is possible, Woollard said, but because of liability and safety issues, the tech is more often used for prevention.
“We are focused more on loss analytics, scoping the problem, understanding the problem, addressing the problem,” he said.
InStore.ai also monitors employee theft or employees giving away products, but Meyer said that he’s still assessing how he wants to handle those instances.
“Is that a write-up? Is that a termination? Do we want to terminate off of an audio recording? It's quite the rabbit hole to go down on some of that stuff right now. Now we're just monitoring it,” he said.
Employee Perspective
Aren’t employees uncomfortable with being recorded? Meyer said there has been little resistance to the technology.
“We've been recording [their audio and video] for years,” he said, referring to security cameras. “I was really worried that they'd have an issue with it. I don't think they really differentiate between what we were doing and what we are doing. We have not had one employee have any issue with it, and they all were made aware of it.”
AI can help management reward employees. For example, while a traditional transaction log can show what a customer bought, it doesn’t show if they added something to their basket because they saw a promo sign or if an engaged cashier told them about it.
“Nothing says if it was a cashier-assisted sale,” Blazensky said, and when employees don’t get recognized, they might not care to continue to stay motivated.
“It’s typically a high school-educated employee getting paid just over minimum wage, gets little training, little recognition and is thinking about quitting,” he said.
Cashier engagement can also be compared to other stores or the industry average, Blazensky said. Retailers can train employees based on what the data shows to be successful.
“I take a lot of pride in feeling like I have a real pulse of everything that's going on in the stores, and I take a lot of pride in being at store level,” Meyer said. “As you grow and you spread yourself out so thin and you have so many demands on your time, you really start to feel another step removed from really what's going on at store level. It's stuff like InStore.ai that allows you to efficiently get a gauge of where things are going.”
It also provides as a safeguard for employees who are falsely accused by customers of doing something wrong.
Sultan is confident that employees will strengthen their skills with the help of InStore.ai, since it can provide models for conversations with customers. Those recordings can then be used for training purposes.
“We have to be able to answer confidently and give examples to our staff,” he said. “Just telling them is different; showing them here and now is another thing.”
Other AI Uses in Retail
Sultan has a background in information technology, computer networking and short films, and he enjoys fusing retail with technology and his creativity.
Fav Trip’s primary marketing channel is YouTube, where the company posts videos busting shoplifters, product reviews and more. Sultan said ChatGPT—a chatbot that uses AI to have human-like conversations and produce text, video and images—helps brainstorm ideas and video titles.
“We're about 1,000 away from getting 100,000 [subscribers on YouTube],” Sultan said. “We’re hoping that, with the help of AI, we can hit that last milestone. [We’re] having fun with it and hoping other retailers find it as a useful tool.”
Sultan also relies on Claude.ai, another artificial intelligence chatbot that can hold conversations.
“It’s like your smart college friend that you bounce ideas with … it won’t have your final answers, but it’s fun having that back and forth,” he said.
Testing out New AI
It’s important to test these technologies out, Sultan said, because “expectations are usually very high, and for new technology, sometimes it's not quite there yet.”
There’s a fine line between AI that brings value to the company and AI that does a task that an employee could have done in the same amount of time, Sultan said. In choosing which AI to use, Sultan suggests evaluating carefully.
“How much time, energy and money are you willing to put in to do this? It gets overwhelming,” Sultan said. “There is a tool being released almost every day.”