The Trust is exploring automation opportunities, recognising that documenting current processes was a crucial first step. This involved a three-stage workflow. First, they conducted interviews with staff members involved in specific processes, recording these meetings via Microsoft Teams. This method allowed the interviewers to focus on actively listening and guiding the conversation, rather than taking extensive notes. The recordings also facilitated an initial process check before involving subject matter experts.
Next, the Trust employed Google AI Studio, utilising its latest language models, to convert the meeting transcripts into detailed process documentation. This was achieved using the following prompt:
“You are an expert business analyst tasked with reviewing a comprehensive transcript from a meeting with two HR specialists.
Your goal is to document the process described in the attached meeting transcript in great detail. This process should be highly logical, step-by-step, and describe every step in minutia.
Carefully read and analyse the entire transcript. Pay close attention to any mentions of processes, procedures, or workflows discussed by the HR specialists.
As you review the transcript, identify the key steps in the process being described. Look for sequential actions, decision points, and any specific tasks or responsibilities mentioned. Document the process in extreme detail, ensuring that you capture every step, no matter how small it may seem.
Your documentation should be highly logical and presented in a clear, step-by-step format. Describe each step in minutia, including any relevant information such as:
Who is responsible for each step
What actions are taken
What tools or systems are used
Any decision points or conditions
Timeframes or deadlines, if mentioned
How information is transferred between steps
Any potential variations or exceptions in the process”
The prompt emphasised the need for a highly logical, step-by-step breakdown, capturing every action, responsibility, and decision point. For example, the prompt instructed the AI to identify who is responsible for each step, what actions are taken, which tools or systems are used, and any decision points or conditions involved.
Finally, the AI-generated process documentation was reviewed and verified by subject matter experts within the Trust to ensure accuracy and completeness. This iterative process, combining AI-driven analysis with human expertise, resulted in highly detailed and reliable process documents suitable for staff training, policy development, or process handover. The Trust found that being focused and specific during staff interviews was key to obtaining high-quality transcripts for AI processing.
Key Learning
Utilising AI for process documentation significantly reduced the time and effort required to create detailed workflows.
Focused staff interviews, recorded for accuracy, combined with precise AI prompting using Google AI Studio’s latest models, proved highly effective.
Risks
Reliance on AI necessitates thorough review by subject matter experts to ensure accuracy.
Data privacy and confidentiality must be carefully considered when recording and transcribing staff interviews, with appropriate consent and security measures in place.
Over-reliance on AI without human oversight could lead to inaccurate or incomplete documentation.
