I gave my AI a decade of my work

This Week, For Us | Issue #005

Jul 24, 2026

One of the things I did when I decided to move Hey Systems onto Substack was go through every asset I had ever created in my business.

And, as someone who has spent far more time building the backend of her business than the frontend (for better or worse) there was a lot.

I created a Markdown file for every lesson in every course. I pulled the transcripts from every video. I added the workshops, frameworks, templates, and resources I have built over the years.

I also added the more than 1,000 emails I have sent to my list over the last decade.

All of it now lives in a context library so that Claude can easily reference, pull, and repurpose things that I’ve created.

Alongside that archive, I have a set of foundation files explaining the business itself: who I help, what I believe, how I work, what I am building…

Because of this, when I work with Claude as a thinking partner, it has the depth of context it needs to support me with high-quality strategic thinking in my business.

I’m not starting each conversation from scratch or asking AI to invent a direction for me.

It can look across years of work, find ideas and patterns I may have forgotten about, and help me see what has stayed consistent as my business has changed.

That has been especially useful as I think about how I am building Hey Systems over the next year.

Last week, I spoke on stage at a convention to a group of entrepreneurs for close to three hours about AI.

And one of the key takeaways I shared was that there are two main things AI needs in terms of training to do an exceptional job of supporting you:

A Context Library

The information AI needs to understand your business: your customers, offers, ideas, decisions, values, voice, constraints, and existing work.

A Process Library

The instructions AI needs to understand how work gets done in your business: your workflows, processes, standards, and preferred ways of approaching recurring tasks.

Capturing the knowledge of your business and the workflows and processes of your business means that, no matter which AI you’re using, you can give it those details and help it work effectively for you.

One tells AI what it needs to know.

The other tells it how you want it to work.

They’re the same two things you would need to hire and train a new employee.

And because you own those libraries, they can move with you. You’re not relying on one particular model, platform, or chat history to remember how your business works.

And luckily, it’s getting easier than ever to give AI the context and processes it needs to get things done for you in your business.

Are these two systems something you’d like to learn more about? Drop a note in the comments and let me know if you want to build this in your business.

Speaking of capturing workflows…

This week, Anthropic added “Record a skill” to the Claude desktop app (paid plans) where you can your record your screen while you do a task, talk through what you’re doing as you go, and Claude turns the recording into a skill it can run again later. (Brock Mesarich does a great job of breaking it down here.)

As a side note, you might also want to check out the /no-ai-slop skill by Peter Yang.