AI and the Potato Industry
Artificial Intelligence for the Potato Industry — Growers, Industry & Partners
Artificial Intelligence for the Potato Industry — Growers, Industry & Partners
Before the workshop we asked what questions you have about AI and how you are (or are not yet) using it. The responses fell into a few broad patterns. On the question side, people asked where AI will benefit growers most and how it can drive—or even learn—best management practices; how AI is applicable in potato storage; how it is being implemented in agriculture and how it can be used in the industry more broadly; and what specific tools exist for disease diagnostics and research. Some of you are working on research involving detection of pathogens using AI, or exploring applications like using AI to predict antibodies for infection in club root. Others said "not yet" or are in a learning phase. Broader themes included how to get started with data and trust it; what ethical or liability issues to consider; how AI might affect the ag industry and complement technologies like CRISPR; and whether copyright or privacy should be a concern. Some are curious about efficiency of AI in products, software, strategy, and industry intelligence.
On the “how you’re using AI” side, the range is wide: many said they are not using it much yet or are holding back to see how it develops; others are experimenting with ChatGPT for clerical work, business writing, email, SOPs, contracts, and meeting notes. Some are already using deep learning for image classification, optical sorting, hyperspectral imaging, or NLP/ML for prediction and research. We don’t list every response here—instead, the Discussion and Pathways sections below reflect what you told us and turn it into themes and next steps.
Across all responses, two main streams show up very clearly. The first is business analytics and workflow: where AI might fit into day-to-day decisions, how to feed it good data, how to trust the results, and how it might reduce administrative burden. The second is AI/ML and machine vision: specific tools for storage management, disease diagnostics, optical sorting, and research questions. Many of you mentioned that you are not using AI much yet, or are “hanging back” to see how the technology develops. Others are experimenting with tools like ChatGPT for clerical work, drafting documents, and exploring data.
As you read this page, you might want to quietly place yourself on that spectrum: from “not using it at all yet” through to “experimenting a bit” and on to “regular user.” The reflections and examples below are designed so that, wherever you sit on that line, you can identify one or two realistic next steps that make sense for your role and operation.
Many of you asked, in different words, the same core question: “Where will AI actually help me?” You are interested in real-world, concrete examples that fit the realities of potato production, storage, and marketing—things you could reasonably try in the next season rather than far-off science fiction. As you move through the workshop, watch for ideas that tie directly to problems you already have on your desk: a decision you keep revisiting, a report that always takes too long, or data that you already collect but do not fully use.
Several responses focused on getting started without being an expert in data or coding. You asked about the “best way to input data” so it can be analyzed, and how to begin even if you are skeptical or feel behind. In practice, that often means beginning with what you already have—field notes, storage logs, grading data, contracts, SOPs, and photos—and using AI to help structure, summarize, or explore patterns, rather than building a complex system from scratch on day one.
You also raised trust, ethics, and privacy. Questions like “Why should I trust AI?”, “Who sees my data?”, “What are the liability issues if the model is wrong?”, and “How do we build a fair data ecosystem across the food chain?” are front of mind. During the sessions, whenever we walk through an example, we will pause to ask: What checks would you put in place? Who should own and control the data? How could this change relationships in the value chain?
Another theme is workflow and roles. Rather than replacing agronomists or managers, many of you are curious about how AI might extend what those roles can do—by drafting SOPs and policies, turning meeting notes into clear next steps, or pulling together information from multiple sources so that field visits and conversations can focus on judgment, experience, and local conditions.
Finally, some of you are already looking ahead to specific tools: AI for potato storage management, image-based disease diagnostics, hyperspectral imaging and optical sorting, robotics at harvest, development of seed cutting technologies, and how AI might sit alongside technologies like CRISPR/Cas. Others raised questions about using AI for forecasting yields, targeting agronomic or disease-management interventions, and even early‑warning signals around financial risk. For you, the workshop and linked pages highlight where these tools are today, where they are still emerging, and what good questions to ask vendors, researchers, and partners if you are considering a pilot.
ROI and resource optimization; yield forecasting; data input and reporting; admin (SOPs, contracts, policies); traceable measurements; trust, ethics, privacy, and data ecosystem; making agronomists more effective.
Tools for potato storage; disease diagnostics from images; optical sorting; deep learning and hyperspectral imaging; AI in research and seed cutting; robotics and automation; which programs are proven in ag.
Many “not using much yet” or testing but hanging back. Others: ChatGPT for clerical work; synthesizing information; data analyses; lead generation and research; NLP/ML for prediction; image-based classification; optical sorters.
The workshop and site are organized around the same two streams that showed up in your feedback. For deeper dives tailored to each stream, see the Feedback sub-pages: Business Processes (workflow, ROI, admin, trust) and Machine Vision (storage, disease ID, optical sorting, research tools).
If you are mainly interested in business analytics and workflow, spend extra time with Business Processes and the Business Analytics & Workflow page and related parts of Session 1. As you do, consider one process you are already responsible for—a regular report, a planning meeting, a storage decision—and ask how AI could help you see patterns sooner, explore “what if” scenarios, or cut down on repetitive drafting work.
If your interest is more in AI/ML and machine vision, see Machine Vision, the AI/ML & Machine Vision page, and the case studies on storage, disease ID, and optical sorting. As you review those, think about where you already collect images or measurements—grading lines, storage inspections, field scouting—and how those could connect to emerging tools. You do not need to build a model yourself; a useful step might simply be learning what questions to ask about accuracy, data requirements, and integration with your current equipment.
For many people, a good way to begin is with a low‑risk first experiment. That could be using an AI tool to summarize a field report, draft an HR policy, explore how different scenarios affect ROI, or turn handwritten notes into a structured checklist. The Exercises page includes a Business Analytics & Workflow exercise that walks through this kind of transformation—from rough notes to clear, traceable output—so you can see what “better data in” looks like in practice.
As you try things, keep returning to questions of trust, ethics, and privacy. Make it a habit to check AI‑generated outputs against your own knowledge, to be explicit about who can see which data, and to think about how shared data might benefit both growers and processors rather than only one part of the chain. The Resources page points to potato‑focused initiatives and media pieces you can use to deepen that conversation with colleagues and partners.
Finally, when you think about AI, try framing it as an extension of your team rather than a replacement. Let it handle more of the routine drafting and pattern‑spotting so that agronomists, managers, and growers can spend more of their time on interpretation, local knowledge, and decision‑making in the field and the storage. Over time, a sequence of small, well‑chosen experiments will do more for your operation than a single, large, risky leap.
Registrant feedback splits into business/workflow themes and AI/ML/machine vision themes. People want practical impact, clear first steps, and guidance on trust and data. The workshop and site content are aligned to both streams, with new sub-pages, exercises, and this Feedback summary so the session stays responsive to what people are asking for.