Utilizing Natural Language Processing to Author Bioanalytical SOPs

We will focus on the process to effectively develop templates and prompts for AI-powered Large Language Models (LLMs) to create standard operating procedures (SOPs) for the Bioanalytical Laboratory. We will present various methods and strategies that have been developed to implement comprehensive templates that ensure the consistency of AI-generated content. The importance of longer prompts will be emphasized and show how they enhance AI understanding and output accuracy. Additionally, we will present benefits of implementing a feedback process, enabling iterative improvements and refining the AI’s performance over time to produce sufficiently detailed SOPs. Methods to ensure the SOPs provide correct information will be presented as without a good template and explicit instructions, all the current major public AI options (ChatGPT, Llama, Gemini) produce errors in text of SOPs. By applying the techniques presented here, we show how AI generated Bioanalytical SOPs can meet high standards of clarity and compliance.