These are a few ideas that I scribbled down on some post-it notes from last week. Sometimes, I'm not at my computer, or things just come into my head while I'm thinking. Post-it notes (pen and paper) are some of my fastest tools.

Notes from 12/16/25 - post-it notes

- Run through some scenarios with the AI agent. See how it is doing.

- While on my hike today, I was thinking about training the AI agent on how to do our style of accounting (following the data through the different processes).

- Train both human and computer

- Document everything and keep it public facing

- I kept thinking about a kid who presented his code project, with a known bug, to get feedback and get advice on how to fix it. That was bold and inspiring. Sadly, I like to have my stuff together and I'm scared of presenting things with bugs or huge known issues. I would rather hide and work them out. I may have to get over that... That may be a challenge! :)

- On the meeting with the AI guys (tonight 12/16/25), Ted was trying to push us to get SOP's, flow charts, processes, good documentation, walk through the reasoning, and even solicit our clients and see if we can get their SOP's and documentation. We, as a company, know that some of our clients have paid big money to get SOP's made for them. I'm not sure how that will go over, but worth asking.

Notes from 12/17/25 - more post-it notes

- On global context... either use an internal API or mirror certain information over to that server. Which way would be the most effective?

- Standardize the output

- I have data in EOT's (elements of time), media/content tables, photo galleries, subs of time, and flex grid. All of these are database tables and have content that I want for global context stuff.

- Create a new mixed mini API - pull from help files, elements of time, subs of time, photo galleries, media/content, and the new global context database table. Dump it all back as a JSON file through the mini API.

- Adilas University - All global context should go there. This was a small break through for me. I have it all over the place. Then I was thinking about where it should go. The thought came up about using the adilas university (site that already exists - with full adilas database functionality) and using that site or corporation. That will work slick and perfect.

- On the media/content, photo galleries, and subs of time stuff - only pull from things that are publicly facing - keep private, private

- On the database side, maybe duplicate or mirror certain pieces - main control will still be from inside the system (based on normal corporations, users, pages, and permissions). Keep the control stuff on the normal corp level. That is already built in. If we need more help, in the adilas university site, we can assign users and give them instructions and permissions where needed.

- When gathering content, look under the my docs folder and my personal folders on my local computer. I often save things to those spots. It may make it easy to go there and scoop things up. There are other folder where I stash things, per person, based on what we are working on. If we are really gathering content, maybe look there. Sweep the floor, turn over rocks, and look under bushes (virtually).

- On the mixed mini API results - maybe be careful for the term adilas - be white label sensitive - we want to pitch and sell our stuff, but part of our plan is allowing rebranding and white labels. We may have to do some global find and replace stuff. Maybe just do it normal first, then circle back around.

- Model the data - do some samples and look at what we want to get back - do some prototyping and mock-ups, just like other database or code projects. Design the outcome or output.