Audyx is now LIVE

This week I officially wrapped up Audyx Beta and everything I needed to get Audyx launched.

In the last month:

·We ran our first full audit with our beta testers and used Voiceform to collect the data—highly recommend them!

·From there, I built 4 Custom GPTs in order to process results, generate summaries, and uncover insights for clients.

This was fun as it was heavy prompt engineering in order to create my own custom AI tools.

For example, my Scale Based GPT takes a file with zero context and automatically creates visuals, a summary, and insight—all while self-auditing its work.

This time was equally exciting and frustrating, as I spent many hours “arguing” with ChatGPT and my tools to debug errors and get things running smoothly.

The most important thing I learned in this time?

Humans are still greatly needed in every element of AI.

At Audyx, every tool includes QA safeguards and I have a dedicated QA GPT.

But! I’ll also be personally reviewing every single client output to ensure accuracy, clarity, and context.

Up next: I recorded a short video to share what’s been happening behind the scenes, why I built Audyx, and where we are headed.

If you know a founder (50-250 employees) who might find this valuable, please pass it along!

Early next week I’ll share the actual deliverables we created during Audyx Beta so you can see what it all looks like.

Michael

PS. I named my main ChatGPT session Pete.

It’s a bit insane, I know. But it’s the main chat I’ve been brainstorming, researching, trouble shooting, and building with. So it felt appropriate—and is also a nod to FP’s first fictional employee when I started FP in 2021.