Artificial intelligence is moving quickly from a curiosity to a practical tool in supply chains, but speakers at the Sosland Purchasing Seminar in Kansas City made clear that the technology’s real value lies less in replacing people than in giving them back time.
Erin Nazetta, an executive adviser for agriculture, commodities and global risk, said AI can accelerate the gathering of information from commodity reports, production lines and customer orders, easing the burden of ...
Continue Reading This Article
Enjoy this article as well as all of our content, including reports, news, tips and more.
By registering or signing into your SRM Today account, you agree to SRM Today's Terms of Use and consent to the processing of your personal information as described in our Privacy Policy.
manual spreadsheet work and allowing buyers and purchasers to spend more time on relationships and commercial judgement. She told the seminar that the pace of adoption is still uneven in food and agriculture, partly because the industry remains highly physical and much of its data is still not digitised.
A poll conducted during her session suggested many companies are still testing the waters. Nearly half of respondents said they were running pilots without committing to a full rollout, while a smaller share had embedded AI into daily work. That split reflects a sector that sees the promise of the technology but is still wrestling with the basics of implementation.
Katherine Parr, senior food and beverage solutions consultant at Aptean, said the most immediate gains are coming in routine tasks such as daily commodity reporting. In her view, work that once required repeated manual collection can now be handled more quickly and consistently, freeing planners to focus on analysis rather than data entry. She pointed to use cases in baking, where AI can help track ingredients such as flour, sugar, cocoa and fats, and assess how changes in protein content or commodity prices may affect yield and formulation costs before they disrupt production.
Parr also said AI can help reduce downtime on production lines by improving the sequence of changeovers, particularly where allergen management is involved. By organising production runs to move from less allergenic items to more sensitive ones, and even taking into account colour or flavour profiles, the technology can help cut washouts and improve efficiency.
Beyond current applications, she said the next major opportunity may be in combining operational data with external signals such as weather and tariffs. That would allow sourcing teams to better anticipate the effect of droughts, crop stress or trade shifts on ingredient availability and pricing.
The discussion also drew a distinction between large language models and agentic AI. Parr described the former as reactive, able to answer prompts and generate summaries, while the latter can carry out multi-step tasks, such as monitoring supplier performance, flagging a problem and routing it for approval. In her view, that difference matters because simple conversational tools are useful for quick questions, but industry-specific systems are more likely to improve forecast accuracy and other business outcomes.
Marc Losito, vice-president of regulatory solutions at FoodChain ID, made a similar case for tools designed around the realities of food supply chains. He said the company’s Scout platform combines a language model with agentic functions so it can read and summarise regulatory and commodity data, then move on to monitoring, flagging exceptions and sending issues to the right people. The system draws on proprietary, partner and public data, including futures prices, weather and geopolitical feeds, to spot early warning signs that may affect sourcing, compliance and fraud risk.
Losito argued that this kind of continuous monitoring can help buyers see problems sooner than they would through manual research alone. He also cited geopolitical disruption and fertilizer markets as examples of how shifts in one part of the supply chain can ripple into grain availability, ingredient costs and even the risk of economically motivated adulteration.
Even so, both speakers stressed that AI is only as useful as the data behind it. Industry surveys cited by Aptean suggest that data quality and access remain the biggest obstacle for most food and beverage decision-makers. That concern echoes wider industry warnings that AI can magnify weak processes, inconsistent records and legacy-system fragmentation if the underlying data architecture is not sound.
Nazetta said the technology can speed up analysis, but it does not remove the need for experienced judgement. Buyers, sellers and hedgers still have to interpret the output, weigh commercial relationships and make the final decision. Parr agreed, saying AI should be used to take on the heavy lifting of gathering and comparing information, while people retain responsibility for confirming recommendations and acting on them.
Losito was blunter. Generic AI, he said, can identify patterns quickly, but it cannot reliably decide what those patterns mean in a regulated, supplier-dependent environment. In his view, any credible deployment in supply chains must keep a human in the loop.
Source: Noah Wire Services
