Enrico SchaeferFouning Attorney, Litigator and Ousdie Counsel - IP, Tech, TRAVERSE LEGAL, PLC · Founding Partner @ TRAVERSE LEGAL, PLC | AI, Blockchain, Social Media, Internet

Unleashing the Power of AI 'Speech-to-Text' Transcription on Your Law Practice.

2 Apr 2024 · 2 min

Watch on YouTube

2:12

In this presentation, Enrico Schaefer demonstrates how to harness AI transcription tools—specifically Fireflies for meeting recordings and Whisper for rapid voice‑to‑text dictation—to capture and upload audio transcripts. He then shows how a custom GPT assistant can analyze those transcripts to surface the "quad‑core foundational elements" used in a training video aimed at prospective client calls. The identified elements include prompt engineering, providing examples, method/strategy, repetitive‑task automation, specialized skills, and the right mindset. Schaefer emphasizes the speed advantage of dictating prompts (under five seconds) versus typing, and illustrates the end‑to‑end workflow of recording, transcribing, uploading, and extracting actionable insights for law‑practice workflows.

Chapters

A big proponent of Fireflies. It's an AI transcription tool. You can have all of your meetings recorded, transcribed, and run AI on them. But you can also download transcripts of any audio file, whether it's recorded in real time or you export the audio and then feed it into this tool. So now I have a downloaded transcript of the class thus far. So what I did is I uploaded that transcript, and then I said, hey, I'm going to upload the transcript of the video that I'm working on. This one is for prospective client calls. I want you to identify each of the quad core foundational elements that we used in this training video. For instance, now prompt engineering is about guiding your AI assistant giving it the appropriate context and instructions it needs to succeed. And one of the most powerful things available to us as prompt engineers is providing examples. The, the data is very clear. If you provide examples to the GPT, its ability to provide better output is improved exponentially. So let's take a look at what I say. Say, for instance, we had to think about method or strategy, what repetitive tasks might be good for building a GPT. We had to use skills on very specific crafting to build this GPT mindset. We had to have the right mindset. And keep in mind, I actually used the Whisper AI tool to dictate this prompt, which I was able to do in under 5 seconds, much faster than typing. And so now, just to make sure that it had the transcript, you could read it, I went ahead and cut and pasted it. I could have uploaded it, I could have cut and pasted it, same thing. Let's see what my assistant, my GPT assistant, did to help me help you understand which of the quad-core foundational elements we use so far in this training video.