Using AI to safely synthesise global focus-group transcripts.
To ensure the clients course reflected real, lived experiences across regions globally, the Learning Experience team ran multiple global focus groups. These sessions explored sensitive and personal topics, with participants joining from different roles, cultures, and time zones.
Over several weeks, this created a large volume of qualitative data that needed careful, respectful analysis before it could be used in learning design.
Across a one-month period, 7 focus-group sessions produced over 9 hours of transcripts. It became clear that insufficient time had been scoped for a Learning Designer to realistically re-read, analyse, and synthesise this volume of material.
Early reviews showed that key themes were emerging across multiple conversations. However, identifying and validating those patterns meant jumping back and forth between seven separate transcripts, increasing cognitive load and review time.
The sessions were open and conversational, transcripts included: small talk, tangents, clarifying questions. While valuable in conversation, this content added significant noise when reviewing transcripts line by line.
After identifying a bottleneck in the project workflow, I proposed a more efficient approach using NotebookLM to support the team in collating key themes from focus-group transcripts faster, whilst protecting timelines and avoiding unplanned scripting costs without losing valuable infomation.
Set-up a shared NotebookLM workspace using real focus-group transcripts.
Ran a live walkthrough with the Learning Designer, showing how targeted prompts surface key themes, contrasts, and patterns.
Explained how AI-assisted summarisation could reduce review time and protect space for thoughtful scripting decisions, using a secure and responsible AI tool.
$1,125
Saved on a single project — with clear potential for further efficiencies across future work.
What was previously a multi-day manual review process was reduced to a few hours, protecting project scope and avoiding unplanned scripting costs. This allowed the Learning Designer to focus on higher-value learning design decisions.
Introducing NotebookLM through practical demonstrations and real project data helped build trust and encourage adoption without adding pressure to the team. This project alone saved approximately $1,125, with clear potential to deliver further efficiencies across future work.