AI Playground (Experimental Feature)

The AI Playground is an experimental feature that allows you to export detailed data for selected lots, pens, or animals into individual JSON files for use with AI tools such as ChatGPT or Claude. Each selected record generates its own file, named after the lot, pen, or animal. These files contain structured data that AI systems can analyze, allowing you to explore your own real feedlot data using natural-language questions and see what kinds of insights may be possible.

This feature is part of our research into how artificial intelligence can be meaningfully integrated into Fusion. Before building a fully integrated AI capability, we need real-world feedback on what is actually useful to feedlot managers and owners. Your experimentation and feedback will directly influence how this evolves into a fully developed feature in a future release.

How the AI Playground Works

When the window opens, you will see two main areas: a feedback section and a save section.

At the top of the window is a feedback field. This allows you to send us comments about your experience as you experiment with the feature. You can share what questions you tried, what responses were helpful, what was unclear, and what you wish you could ask. Clicking Send submits your comments directly to us. You can send feedback anytime — not just once — and even short notes are extremely valuable in helping us refine this feature.

Below the feedback area is the Save Folder section. By default, files will be saved into a "Fusion AI" folder on your Desktop, but you can click the … button to choose a different location. The window will indicate how many files will be created based on your current selection. Clicking Save generates one JSON file per selected record. Clicking Cancel closes the window without creating files.

Using the Exported Files in AI Tools

Once the files are created, you can upload them into AI tools such as ChatGPT or Claude. Most modern AI systems allow you to attach files directly to a conversation. After uploading one or more JSON files, you can begin asking questions about the data.

You may upload a single file to analyze one lot, pen, or animal or multiple files to compare performance across records. Because the data is structured, the AI can summarize performance, identify patterns, compare records, and highlight unusual results.

Example Questions to Try

For individual lots:

• "Summarize the performance of this lot."
• "What factors most affected profitability?"
• "How did death loss impact cost of gain?"
• "Identify anything unusual in this lot."

For comparisons:

• "Compare these two lots and explain the main differences."
• "Which lot was more profitable and why?"
• "What management factors might explain the performance difference?"

For pens or animals:

• "Are there patterns in feed efficiency across these pens?"
• "Summarize this animal’s treatment history and performance?"
• "Does this animal differ significantly from the group average?"

You can also experiment with broader prompts such as asking what a feedlot manager or owner should focus on, what follow-up questions should be asked, or what risks might be present in the data.

This Feature Is Experimental

The AI Playground is intentionally labeled as experimental. We are exploring important questions such as:

• What types of AI-driven insights are truly helpful in day-to-day feedlot management?
• What data needs to be included?
• What analyses should be automated inside Fusion?
• How should AI be integrated into existing workflows?

The answers to these questions depend on real-world usage. That is why your participation matters. We encourage you to take time to experiment with this feature and to share your feedback regularly using the built-in feedback field. Try different types of questions. Upload multiple records and compare them. Push the boundaries of what you think might be possible.

Getting Here

You can open this window by clicking the AI in the Lot Center, Pens List, or Animals List windows.