From the fields and orchards that grow produce to the packers and shippers who send truckloads of fruits and vegetables to a network of distribution centers, retail stores and foodservice operations, the supply chain has long relied on technology.
As artificial intelligence (AI) and smart devices offer new tools to companies throughout the supply chain, relying on the same decision-making process to assess and implement technology can lead to failure. While costs, scalability and life-cycle analyses are still critical, industry and retail experts say the decision-making process shaping technology procurement has changed greatly in recent years.
Global tech media, data and marketing services firm Foundry surveys vendors on changes in the way their customers make decisions. According to Foundry’s 2025 Role & Influence of the Technology Decision-Maker report, tech purchases are being made by an expanding cross-section of company employees, and vendors are tailoring pitches for specific roles at those companies. At the same time, nearly all information technology decision makers surveyed say AI has had some influence on the buying process — aiding research, product comparison and cost analysis.
The AI Integration Problem
Throwing a wider net for input on the process doesn’t end there: an organization that rolls out AI as a separate tool for departments to use exclusively for their own needs is missing the potential efficiencies of an overview approach.
Gary Hawkins, CEO of Retail Mindsteps, a consultancy advising grocery and consumer-goods retailers, as well as technology providers, says retailers often err by applying AI to optimize isolated functions such as product assortment, demand forecasting and pricing, instead of using it to reshape decision-making across the organization.
Marketing, merchandising, procurement and other areas continue to operate in separate silos, he notes.
“And each of those functional areas has systems — data that they have built up over the years to do what they need to do,” says Hawkins. “You’ve got data broken up across disparate systems across the enterprise.”
Companies must pool the information, “weaving a data fabric” for a more comprehensive view of the overall operation, he says, bolstered by agentic AI rather than a generative AI system. Generative AI creates text, images and other content based on human prompts, while agentic AI is autonomous and can perform tasks by planning and reasoning.
These AI agents can focus on tasks ranging from launching a new product to managing a fresh produce inventory, says Hawkins.
“What we’re talking about is really the creation of an enterprise that has a much clearer insight into their business because of all the data unification,” he says. “It is becoming increasingly automated — I like to use the word ‘bionic,’ you know, a fusion of computer software and human beings to operate.”
Bruce Peterson, president and founder of retail consultancy Peterson Insights Inc. and a 50-year veteran of the fresh produce industry, says precise and timely communication with those who will use the technology must be the foundation for a successful rollout.
“And often, it’s not just a matter of internal communication, but how will a new change affect suppliers, consumers and other groups that this would impact?” says Peterson.
Vonnie Estes, vice president of innovation at the International Fresh Produce Association (IFPA), says productivity tools like email and document management — a common business use of AI — remain helpful in everyday use, but the opportunity ahead is full-chain adoption. For example, using intelligent cold-chain sensors that don’t just detect problems but recommend actions to reduce shrink and food waste.
Culture, Cost and the ROI Test
For all of AI’s potential, as a standalone platform or in devices that ensure quality on packing lines or allow machines to determine a weed from a vegetable plant, it’s reasonable to assume a technology’s complexity can derail operations from time to time.
Technology itself, however, is rarely the cause when a system fails to operate properly, says Estes. That’s applicable across all forms of technology in the supply chain.
A common mistake in AI adoption is focusing on the technology itself and not its intended goals in organizations, notes Estes, who specializes in technology adoption throughout the supply chain, helping companies achieve practical and scalable adoption of systems.
“A lot of times, people say, ‘We’ve got this software, this program, this AI, and we’re just going to bolt it onto these legacy systems that we already have in our manufacturing plant, in our sorters or anyplace in the supply chain,’” she says.
Hawkins agrees. “While it sounds like it can be headache-causing, the technology piece of it, in my mind, is almost the easiest piece,” he says. “The harder piece is the organizational change, the cultural change that has to take place, to really take advantage of what new technology has to offer.”
The key is to focus on the intended goal, whether it’s to solve a problem, increase efficiency or extend the quality and shelf life of fresh produce, says Estes.
Another critical step is retooling workflow management, ensuring workers are trained, and adopting a top-to-bottom “culture of technology,” she says.
“If you have an organization that isn’t used to adopting new technology or you don’t have a learning organization, and you try to force anything in…you’re not going to get adoption and have success,” asserts Estes.
Peterson says that when his clients discuss technology, it’s in relation to three areas: understanding and analyzing consumer behavior, supply chain transparency, and warehousing and logistics. And without fail, AI dominates discussions about technology. Peterson left Walmart Stores Inc. as the senior vice president of perishables in 2007 after 17 years in different roles during the retailer’s massive expansion across the country.
Peterson, Estes and Hawkins point to Walmart’s long-standing embrace of technology as a retail leader. Walmart’s corporate website identifies the company as a “people-led, tech-powered omnichannel retailer,” which is demonstrated by its increasingly personalized shopping experience, backed by massive investments in technology. This includes early adoption of digital ordering and other non-traditional shopping methods that predate the COVID pandemic, and automated distribution centers that rapidly fill orders based on a particular store’s layout, prioritizing placement of each item on the pallet.
When Peterson was at Walmart, he helped roll out radio-frequency identification tags to track fresh produce shipments.
“Walmart has a foundational cultural belief in keeping resistance to change low,” he says, arguing that the retailer has historically been quicker than most to trial and implement new systems. He credits its founder, Sam Walton, with setting that tone: “He never wanted the company to get too comfortable with what we were doing.”
Hawkins attended the National Grocers Association Show in Las Vegas in February 2026, where AI was discussed in many sessions. It became clear to him that everyone has an idea of what AI is, but few people understand how to apply it to their organization.
“With that understanding, you can step back and say ‘OK, where could I begin to apply this so that I benefit?’” says Hawkins.
A prominent misnomer about utilizing AI is that companies will save money by reducing their workforce. “You can’t save your way to prosperity,” he says.
Avoiding the ‘Valley of Death’
At the beginning of the rise of on-farm machines, from drones to autonomous vehicles, a common refrain from specialty crop growers was that technology providers failed to fully understand their needs and the obstacles to bringing technology to their fields.
That has changed, with numerous tech providers working with growers to demonstrate autonomous apple harvesters in Washington, strawberry harvesters in California and Florida and herbicide-replacing weeders in numerous locations.
Walt Duflock, senior vice president of innovation at Western Growers, says early adopters of technology have three main concerns: confirmation via on-farm trials that the equipment delivers on outcomes, economic data that shows a three-year return on investment (ROI), and how it integrates with farm operations and what support plan is available for the equipment.
“Most growers have scars from new tech that got purchased and never got fully integrated into their operation to deliver the targeted ROI,” says Duflock. “Every startup needs to understand that the past generation of startups creates the scars that the next generation of startups needs to play through to get to commercialization and scale.”
The grower-tech company partnerships have decreased the so-called “valley of death,” says Estes, referring to products that work well in simulated trials, only to fail when scaled for real-world situations. This expensive failure of technology at commercial operations is tied to scalability issues as costs run rampant, operational constraints disrupt farm work or infrastructure needs doom startups, she says.
Companies have to think not only about their current operations, but also future needs and plans. Will packing line equipment need upgrading or be made obsolete? Will it cost more to do the same work? Will it lead to a decrease in product quality?
“You really have to think about all of those things,” notes Estes.
From Field Trials to Scalable Adoption
Any notion that growers are averse to change is a fallacy. They are business owners that must have faith that new tech will be right for their operation before adoption.
California is ground zero for tech companies refining farm equipment in the hope of serving specialty crop growers.
Western Growers, which represents growers in California, Arizona, Colorado and New Mexico, has ramped up its technology resources in the past decade. The Western Growers Center for Innovation & Technology, which opened in 2015, provides on-farm incubator options that reduce costs and time to bring a viable solution to market. John Deere and Reservoir Farms are partners with Western Growers in the initiative.
Venture capital firm Reservoir launched two Reservoir Farms sites in 2025. One is in Salinas, California, on land leased from Tanimura & Antle, with partners including Driscoll’s and Taylor Farms. Another, in Sonoma County, California, focuses on grape production. The farms will host ag-tech startups testing equipment in fields and vineyards.
“The fanciest tech in the world with the shiniest buzzword-compliant marketing will not be (adopted) if the economics for the grower don’t work,” he says. “In robotics, you have to deliver a solution that can improve the grower’s overall cost.”
Western Growers conducts case studies on the on-farm feasibility of new technology. For example, the organization studied the performance of Carbon Robotics’ LaserWeeder on eight varieties of high-density organic leafy greens. A two-season comparison on 3,200 acres showed weeding by hand with H-2A workers cost $2.1 million, and the LaserWeeder and a smaller labor crew doing the same work cost $1.3 million, says Duflock. That result demonstrates how growers can justify buying a LaserWeeder, which can cost up to $1.4 million, with tech support provided.
“In all cases, the grower is willing to test and help measure, and as long as the grower economics are met during the trials, they will embrace new technology and support the operational integration needed,” he says. “Robotics are being purchased by WG members across several specialty crop types.”
Economic factors that weigh in on growers’ decisions to adopt technology include metrics such as how much “downtime” occurs while repairs are handled, and if there will be any scalability issues.
“Scalability and operability are very important — up-time and run time are metrics growers will focus on because of the time-sensitive nature of a lot of these operations,” says Duflock. “If the machine goes down regularly and they need to hire a hand crew with little to no notice, it’s often stressful to find last-minute labor crews, and they usually come with a higher cost than when you plan them in advance.”
Breaking Legacy Norms
As technology transforms the ability to streamline operations and provide the industry tools to make better — and faster — decisions, Hawkins says organizations must rethink their approach to tech decisions.
“One of the problems I see is that companies go back to the legacy solution providers they’ve worked with for the last 5, 10, 20 years, and all that does is simply reinforce the silos,” he says. “It’s not going to get the company to where it needs to go.”
From the fields to the office, as tech tools evolve at a faster pace, an inclusive process can unlock ideas that lead to better performance. An understanding that technology risk is organizational and economic, and not just technical, is critical to starting the process.
