Prompting guide
Getting better answers from AI
Have you ever tried using ChatMDR or ChatFDA, only to wonder if you were missing something? One key to getting the best outcome is a method known as “prompting.” This approach shapes how we communicate our requests to these powerful systems.
Prompting is telling the AI model what we want
Maarten ter Mors, CTO of ChatMDR, has shared some handy tips on framing prompts. One of these insights could completely shift how you use these systems.
“Prompting is basically telling the AI model what we want,” Maarten said. Unlike old-school search engines that just match words, this process calls for deeper detail. For instance, if you’re looking into brachytherapy device classification, you might ask, “What is the device classification of a brachytherapy applicator?” That single shift to a complete sentence gives the AI model a more concrete sense of context, often resulting in more accurate answers.
But there’s more to it than that first question. Tools like ChatMDR thrive on back-and-forth interaction. “You’re really having a conversation here,” Maarten pointed out. You can double-check facts, ask for clarifications, and nudge the model to rethink if something feels off. This give-and-take helps you refine your final answer and explore more detailed discussions.
There’s more to it than the first prompt. You’re really having a conversation here
Maarten even demonstrated a fun angle by creating a podcast script starring fictional hosts, showing how flexible these models can be. Whether you’re tackling medical device regulations or just playing with creative ideas, a well-structured prompt can truly make the difference. Ready to give it a try?