AI tools are becoming part of everyday life. They are used to answer questions, summarize information, and support decision-making. These tools work based on the instructions provided by the user, known as “prompts.” The quality of the answers from AI depends heavily on the quality of the prompt. If a question is asked without a well-structured prompt, the response may be less accurate because the AI may draw from a mix of reliable and unreliable sources.
Farmers are starting to use AI tools to support decisions related to soil and crop management. Because of this, it is very important to design prompts that guide the AI to use more reliable sources of information. In particular, it is important to use regional information and avoid commercial content, as well as unverified information from blogs, forums, and general websites, to improve the reliability of the answers.
With this in mind, we developed a suggested prompt that can help North Carolina farmers get more reliable information from AI tools. This prompt restricts the AI to use information from Cooperative Extension programs at land-grant universities. It also prioritizes locally generated recommendations, which helps capture differences in agricultural practices across U.S. regions. While the prompt was developed with soil fertility and weed science questions in mind, it could be adjusted and used for other areas of agriculture.
How to use it:
- Copy the prompt below and paste it into the AI tool.
Prompt: Conduct a web search restricted exclusively to Cooperative Extension and Extension Service publications from U.S. land-grant universities. Use Cooperative Extension publications (fact sheets, bulletins, crop guides, and Extension-affiliated applied research reports from U.S. land-grant universities). Give priority to recommendations developed under conditions most relevant to North Carolina using the following hierarchy: North Carolina, neighboring states in the Southeast, southern U.S. states, Midwest/Corn Belt states. When recommendations differ across regions, explain likely agronomic reasons (soil type, climate, management system). Prioritize the most recent 10–15 years unless older foundational guidance is still widely cited. For major findings or recommendations, indicate the following: university/Extension system, state, and whether the recommendation appears region-specific or broadly consistent across regions. Include citations and links for the sources referenced. If Extension sources are not available for a point, explicitly state ‘no Extension source found’ rather than substituting other sources. Exclude commercial, industry-sponsored, non-peer-reviewed, blog, forum, Reddit, Wikipedia, and AI-generated sources. Summarize current Extension recommendations, areas of agreement/disagreement, and practical management implications regarding the following topic or question “[INSERT YOUR QUESTION HERE]”
2. In the prompt, replace “[INSERT YOUR QUESTION HERE]” with your question or request, keeping the quotation marks. Try to include “North Carolina” in your question to ensure locally relevant information.
Example questions to include in the prompt:
- “Is it worth using starter phosphorus for corn in North Carolina?”
- “What are the most common soil fertility issues for corn and soybean production in North Carolina?”
- “How to adjust fertilizer decisions for best return on investment in North Carolina?”
- “How to manage Palmer amaranth in soybean fields in North Carolina?”
- “What are the best strategies to avoid herbicide resistance in row-crop fields in North Carolina?”
3. Submit the prompt to the AI and review the response.
4. If necessary, ask follow-up questions to refine or improve the answer from the AI tool. If you are unsure about the reliability or applicability of the information provided, consult your local Extension agent or Extension specialist for further guidance.
We tested the prompt using three popular AI tools with the questions listed, and the responses were generally reliable. We consider AI a valuable tool to help farmers find and summarize information. However, these tools are not error-free, and all responses should be carefully reviewed and evaluated. Before adopting any major management decision based on AI-generated information, we recommend consulting your local Extension agent or specialist to verify the reliability and applicability of the recommendations provided.
