Case Study - Saving Time with Automated Call Summaries
THE BRIEF
An investment business uses an app to record the transcripts of all their sales calls. But the transcripts contain a lot of irrelevant information, and the process of cleaning and summarising them is time-consuming.
The business wanted to automate the cleaning and summarising of the transcripts using ChatGPT, to save time and improve efficiency.
THE RESULT
- Hours of manual time and effort saved
- The ChatGPT prompts are easy to update without any coding or technical knowledge required
"We were interested in building a custom automation flow to improve the efficiency of our meeting transcript process, which had previously consumed a lot of our time. Greg came highly recommended, and did not disappoint. His prompt and clear communication, excellent technical skills, and keen attention to detail impressed us. And his automation flow has already saved us hours of laborious effort! We definitely plan to use his services again in the future"
Michael Poulsen, Canopy Investors
THE CUSTOM AUTOMATION
The entire process is a "Cloud Flow" in Microsoft Power Automate. This was very important to the customer, as they wanted to keep all their data within the Microsoft 365 ecosystem.
The process is triggered by someone saving the call transcript in a specific OneDrive folder. The flow then uses the ChatGPT API to clean and summarise the transcript, and then saves the cleaned and summarised transcript back to OneDrive.
Importantly, the flow also splits the transcript into sections, and sends each section to the ChatGPT API separately. This ensures that the flow doesn't hit the token limit of the ChatGPT API and avoids any "laziness" from ChatGPT.
The ChatGPT API is also built as a custom connector in Power Automate, so it can be easily re-used and customised for other automation flows.
