David Jenyns demos Process Pal, a custom GPT that turns meeting transcripts into structured process docs with titles, descriptions, triggers, and steps. It's fast but needs human review to catch AI mistakes.
I talked about this idea of custom GPTs and we are training different bots to do different things. So I've been working on one we call Process pal. So it's systemology.com forward/process pal and you'll be able to get to that. I'm just going to show you a super quick example of this.
So this is process pal here, systemology.com process pal you get here I've trained it on a bunch of data, but here I'll click my first button to get started and it says please ask me to upload a transcript for you to create a process from. That is great. Okay, so say here is the process attach. Now one of the examples I'll share with you is the one Simon Kelly is capturing our podcast outreach process. I'm going to attach that. That's just the transcription of the extraction call that you will have access to for our outreach process.
And again, I've pre trained this bot and give it a set of instructions and training in the background for you. And it goes ahead and identifies the title, the description, the trigger, the key steps. And it does that by reviewing the transcript. Again, this is infinitely easier than it used to be. And previously it was worth the time and effort to invest in documenting. Now this just makes it a million times quicker. Then it's just a matter of copying and pasting that into.
Your documentation management system like System Hub and then starting to add the details. This is where the human comes in after the fact and you then start to link to subsystems and extra resources and highlighting and just making that extra get those extra tweaks. At the moment, there is no replacement for the human review. The fact is AI will make stuff up when it doesn't know the answer, it just hallucinates the answer. And you can't just 100% rely on the AI. It's going to guess that's all AI does. It goes based on everything that I know. I think the next most logical step or word right now is this because you haven't told me what it is. Again, this comes back to what James was saying on why the extraction is so important. The quality of the output from the AI is so dependent on the quality of that input. And then we just have the human after the fact step in and start to do the polish. The way that I think about it is AI is a little bit like an Iron man suit. You still have the human in the middle of it, but the Iron man suit just enables you to do so much more. It's empowering great team members. It's not replacing great team members. You're getting more done with less because you've got the Iron man suit that you get to put on.
Just a heads up, that clip was taken from a session in our Systems Champion Certification Program, which is the world's first and only program for systems champions to to help you build a systems culture and develop an AI first approach and roll it out inside your organization. To find out more, head over to systemology.com systemschand and I'll see you in.
Remember me? Subscribe To My Channel For More Free Business Systems Content 📈
