AI Made EZ: Using Transcripts to Eliminate Rote Work

(4 Min read)

Of all the basic uses of AI, working with transcripts is one of the easiest and most practical. Popular video meeting tools (Zoom, Teams, Meet) all include built-in transcription, and there are plenty of dedicated transcription tools as well (Otter, GoTranscript, Rev).

𝗔𝗻 𝗘𝘅𝗮𝗺𝗽𝗹𝗲

I recently conducted a requirements interview with a customer who identified ten use cases for AI. During the session, I systematically covered each use case, with clear, vocal statements, such as: (1) use case number, (2) name, (3) problem statement, (4) current process, (5) process owner, (6) related business systems, and (7) any additional considerations. For example, when completing Use Case #1, I explicitly cued the transition: "If you have no other considerations for Use Case #1, let’s move to Use Case #2, HR Onboarding. What's the problem statement?”

𝗧𝗵𝗲 𝗧𝗿𝗮𝗻𝘀𝗰𝗿𝗶𝗽𝘁 𝗕𝗲𝗻𝗲𝗳𝗶𝘁

My customer uses Tactiq for transcription. I imported the transcript into ChatGPT along with a structured prompt (see below). In seconds, I had a clear, structured Use Case Outline that would have taken me hours to manually transcribe and organize from audio alone. Another major advantage: I took very few notes during the interview, allowing me to genuinely listen and follow up intuitively on subtle clues and comments—“Excuse me, what did you mean by…?”

𝗣𝗿𝗮𝗰𝘁𝗶𝗰𝗮𝗹 𝗨𝘀𝗲𝘀 𝗼𝗳 𝗧𝗿𝗮𝗻𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻𝘀

Transcripts are incredibly versatile. In business, they're great for capturing meeting notes, summarizing follow-up tasks, or confirming next steps. Sales teams use transcripts to assess prospect interest and sentiment. Tech support uses them to monitor customer satisfaction. Doctors rely on transcription to focus more closely on their patients instead of extensive note taking.

• 𝗜𝗡 𝗖𝗟𝗢𝗦𝗜𝗡𝗚

Using transcripts with AI is one of the simplest, most practical ways to reduce manual workload. If you're unsure how to start, just open an AI chat and ask it to help you identify ways to use transcription in your daily work. Give it a try—it will eliminate some of your rote manual work associated with taking notes during meetings.

___________________________________________

𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗱 𝗖𝗼𝗻𝗰𝗲𝗽𝘁𝘀

To make things interesting, I tested two different ChatGPT models by importing the transcript into ChatGPT 𝗼𝟯-𝗺𝗶𝗻𝗶-𝗵𝗶𝗴𝗵 (a reasoning model) and 𝗚𝗣𝗧-𝟰.𝟱 (a traditional GPT model). Both models received the same prompt: “Use the attached transcript to organize the seven attributes of each use case into a clear outline. Additionally, since I didn't explicitly ask for the AI goal during the interview, please infer the customer's intended AI goal for each case.”

𝗖𝗼𝗺𝗽𝗮𝗿𝗶𝗻𝗴 𝘁𝗵𝗲 𝗢𝘂𝘁𝗽𝘂𝘁𝘀

𝗼𝟯-𝗺𝗶𝗻𝗶-𝗵𝗶𝗴𝗵 produced a noticeably clearer and more in-depth outline, while 𝗚𝗣𝗧-𝟰.𝟱 captured several nuanced details. Finally, I asked a reasoning model to compare the two outputs and highlight key differences.

𝗛𝘂𝗺𝗮𝗻-𝗶𝗻-𝘁𝗵𝗲-𝗟𝗼𝗼𝗽

Starting from the 𝗼𝟯-𝗺𝗶𝗻𝗶-𝗵𝗶𝗴𝗵 outline, I used insights from the comparison to manually incorporate important nuances from 𝗚𝗣𝗧-𝟰.𝟱. The result was a polished, accurate outline, shared with the customer in a Google Doc, to confirm their AI use case requirements.