Imagine, for a moment, that you are standing not in a gleaming Silicon Valley office but in the backroom of an American grocery store or a decades-old steel processing facility. There are no sleek product demos or lofty startup valuations, just chatter of local shoppers or the hum of machinery. And yet, it is here, in these grounded, often-overlooked corners of the American economy, that the real story of AI is unfolding.
Here, we share the experiences and thoughts of two leaders: Luke Kotara, who helps run Berkot’s Foods, a regional grocery chain, and Aaron Dallek, CEO of both MD Metals and the AI-driven startup DMaterial. Both companies represent two divergent paths, but their relationship to AI is surprisingly similar. In a sense, cautious, incremental, and rooted in real-world returns.
For Berkot’s Foods, AI is quietly but meaningfully changing the way a small grocery chain operates. Luke explains that they have started with simple automation, such as answering phones and reducing labor costs. Moreover, Berkot’s Foods is moving toward AI price optimization by utilizing historical sales data to determine the most profitable price points for various items.
“We’re using it for automatic answering of telephones, saving labor there,” he says. “Now we’re looking into price optimization, where it knows the movement at certain prices and where we can sell the most at the most profit.”
Meanwhile, in the world of steel coils and flat bars, Aaron’s team at MD Metals gave their office staff access to Microsoft’s Copilot. The verdict? It’s helpful, but not revolutionary. “I don’t know how much efficiency they’re really getting out of it,” he admits. “But I think there’s definitely some.”
Where MD Metals hits a wall is in legacy infrastructure. Their proprietary enterprise resource planning system has no AI integration. Their equipment? Largely from the 1980s or 90s. Integrating AI into this type of environment is not a question of potential. It is one of practicality.
The most visible face of automation or AI in grocery stores? Luke shares that Berkot’s is piloting a new AI-powered shopping cart system from Instacart. These carts identify products as customers place them inside, auto-calculate totals, and let shoppers skip the checkout line. “We’re going to roll that out in the first quarter of 2025,” Luke says.
Another frontier? AI-driven loyalty programs. When customers go missing, Berkot’s digital coupons lure them back based on past preferences. “It knows when a customer comes in, shops, and what they like,” Luke explains. “If they don’t come back, it sends coupons for those items.”
Aaron’s ambitions at DMaterial take automation even further. He envisions mobile manufacturing facilities, equipped with humanoid robots and additive manufacturing equipment, that can be shipped to remote locations. “You could put five, six humanoid robots, some RF welders, and some other additive manufacturing equipment into a container and ship it off to the front lines of a war,” he says. “Now robots cost US$30,000 for one, but in the near future prices will go down.”
For both executives, the future of enterprise AI adoption is not some abstract promise. It is a series of logical steps that start with real use cases.
One recurring theme across the conversation was the mismatch between AI hype and AI reality. Amazon’s Go stores, often hailed as the future of retail, have been scaled back. Luke sees why: “The size of the stores are too big, and the infrastructure is expensive.”
Instead, his team focuses on what works now. “We can’t afford that,” he says bluntly. “We’re working with technology we can deploy store by store, cart by cart.”
Aaron agrees: “I think tech adoption works best in baby steps. Unless you have something that’s truly 10x better, it doesn’t justify the cost. As business owners, we’re looking for ROI. You try something small, see a win, then build on that. Say a solution saves you $10,000 a year. Great. Let’s do five more like it. Maybe two smart carts per store becomes 50 carts. Then digitize shelf pricing. It’s step-by-step.” In other words, you get a win and then go get another.
This approach reflects a core truth: most businesses do not want AI revolutions. They want ROI. They want cost savings, and above all, businesses want dependability.
For Aaron, the next wave is clear: “Quantum is maybe two to four years away from being where AI is today,” he explains. With exponentially more computing power, quantum could drive better predictions in everything from medicine to materials science. It is a long play, but a calculated one.
Aaron also sees a macro shift underway when asked about reshoring. As global supply chains wobble and China’s industrial dominance looks increasingly fragile, Aaron believes US manufacturing will return, but only if automation makes it cost-effective.
“The only way you’re going to do that… is through AI productivity, efficiencies, and automation,” he says. “And that’s not robots taking people’s jobs. That’s just allowing businesses to make things cheap enough to be profitable.”
Luke, for his part, is focused on a more immediate future. Online grocery delivery, curbside pickup, even drone delivery – they are all on his radar. “People don’t want to spend time shopping,” he says. “It’s costing us more labor.” The question is no longer if AI will play a role in grocery, it’s how deeply and how soon.”
The One Big Beautiful Bill (BBB) is reshaping how industry leaders think about innovation, investment,…
A few years back, the only people who could train computers to make predictions or…
EV buyers are no longer loyal to brands. Instead, they are loyal to performance. This…
With nearly 40% of global plastic waste originating from packaging, reducing this waste has become…
The cement industry is the second-largest contributor to global CO2 emissions, lagging only behind the…
Innovation does not happen in isolation. It thrives at the intersection of disciplines, industries, and…