Recording notes from 12/8/25 to 12/15/25. Also spent some time editing AI agent chat logs.

More AI agent training. See element of time # 12542 in the shop for all of the documentation (saved chat logs and starting global context files).

Quick edit for a client and reformatting some rules and regulations for their in-line disclaimer (inside of their corp-wide settings). Sent an email off to let them know that it was finished.

Working with Bryan. Going over one-to-many relationships for parts and items. He has some special pricing issues and questions for one of his clients. We talked about existing options and where we want to go with things. Currently, we have parent to sub items, parent to flex grid tie-ins, or smart group rules and assignments. These are all special one-to-many pricing options.

Going over which one would work best and what we need to change to make that happen. For his current scenario, we spent some time talking about the rules and assignments for smart group buttons. Currently, those rules are based on quantities. We were talking about doing some rules or rule types for sizes or some other characteristic. Good discussion.

Planning and doing some AI agent training for the adilas private model.

See attached for a starting place or file for the Nxtlinq AI agent. This is what it was able to gather to date. I made a few changes... See also the chat log for the new training session.

GoToMeeting session with Steve. We were talking about natural language processing (NLP). API issues (with our current model). Trying to overcome some memory problems and help it use our mini API's. Behind the scenes, the LLM's are getting faster, better, and cheaper. Going over models and options. Lots of talks on different subjects. Talking about API's and where things are heading. Information is becoming more and more easily to get to. Steve is seeing an increase in 3rd party AI tools. People and companies wanting to build out an integration with their AI tools and features. Steve and I have done a ton of 3rd party integrations... We don't really want to keep relearning the same lessons. Currently, he is saying no, vs yes. We've got to get our product fully to market before we do a bunch of new 3rd party solutions.

Recording notes. Prep for a meeting with Steve. We have been working on this AI agent stuff since August of 2025.

Here are a few elements of time that have some plans for where we are heading and wanting to do:

EOT # 12286 - Normal element of time plus media/content - more advanced prompts and tools

EOT # 12392 - aliases and stringing multiple prompts together - interface ideas - deeper prompts and navigation

EOT # 12497 - ideas on training

EOT # 12527 - other plans and global context

Bryan and I met up to look over progress for the restaurant and simple food service functionality. Bryan Dayton has been helping to chip away at this for the past year or so. We did a small review session and took a bunch of notes. Bryan then had the idea to take the new notes and add them to the old notes. We ran them through ChatGPT and it did a great job summarizing things up. We then read over those notes and talked about next steps. Here is the link to our small discussion.

https://chatgpt.com/share/693c91f3-a6bc-8007-a331-7ea33d2b84f2 - notes from Bryan and I from ChatGPT - restaurant and simple food service functionality

After going over plans and notes, we switched briefly over to some other projects. We will meet again on Monday to get into some other projects. We also briefly looked at some hours that have been recorded over on the ship B (adilas shop) side. Both Bryan and I (and others) have been banking some hours on projects that don't have full funding (yet) but we know that they will be needed.

Reinforce
Recruit
Israel (Zion)
Babylon (World)
Reveal & Interpret

---- Notes from a brainstorming and planning session (see attached)

So, just for fun, I had ChatGPT convert my notes into text (OCR style). Here are my notes in a typed form.

- Working on some planning to help with global context & AI/Human training

- I've been gathering some post-it notes for about a week. The electricity is out so I went over to my in-laws house. I'm converting notes from my post-it notes into notes in this notebook.

- 3 notes from 12/6/25: (see below)

Teach the AI agent how to delete or make things inactive - We had some chat messages where a user asked & the AI agent didn't say the right things. We just need to help it out a bit. Coaching...

Teach it how to get to stats & totals. The user asked some basic questions & the AI agent sent them to a decent/okay page, but it could have been better. Once again, more training

Fix & fill in the gaps

------------------

I had a couple of papers (post-it note style) from 12/8/25. Here are those notes:

- I have been thinking about stuff. I will be laying in bed & my mind is going & racing along.

- Make a plan... what tools & options do I have? Use what fits for each situation. It is okay to mix it up & adapt to what works.

- Plan as if for years - I can see some short patches and fixes - plan as if for years, meaning what is it that really needs to be done?

- I have this card & it talks about doing things "Glacier Style" - Carve your own path, go slow, channel your strengths, clear the way for others, stay cool, avoid meltdowns - I really liked that!

- Feed in the glossary - take your time & even add more as needed.

- Look at the prior plans - I've been working on this for months - Sometimes, I get going so fast, I start skipping things

- Work with what you have - that is a great base - you can always make more.

- Use one AI model to train the other ... That could be very useful. Harness the strengths of each model.

- Use voice to text to help with editing - where it fits

- Help the system learn - Take some of the chats & show the request, the response, and any suggestions - maybe help it by giving the answers a rating - just an idea

- Offering to save something for the AI agent - giving it some memory - How would that work? What would that look like

I took a few more notes on 12/10/25 - see below

- Build a small chunk specific tool - include meta data options

- Think in terms of small templates - I want this, this, & that for X - I want this, this, & that for Y - Try to stick to it - for consistency

- Spend some time training the little one - meaning the private AI agent for adilas

I also had some more ideas last night in the middle of the night, and early this morning on 12/11/25 - see below

- On the glossary items - add some terms like: AI, AI agent, AI context window, big data, BI or business intelligence, Knox report, mix, spread, delete, status, Nxtlinq, global context, meta data, vendor credit, special accounts, loyalty points, gift card, threshold, mini scan cart, refund, CardPointe, Clover, go mode, AI quick search, signature, e-signature, printer friendly, bulk add, bulk edit, bulk transfer, AI tools, cash counter, fulfillment, AI image generator, touch or touch screen, modal, retail, dashboard, flex attribute, KPI or key performance indicators, due date, tiered pricing, vendor catalog, delivery, age verification, view only, cultivation, metrc, percentage calculator, sales and profit, Adi, harvest, plant tags, data table, label builder, fracture, chart of accounts, reflexive flex grid, direct deposit, parent/child, try storming, and half baked

- When adding global content what about default sort orders for tools & glossary items? Expanded topics, etc. see next page for a huge list of what we may want to store

- Build a new tool for the AI agent to get at the data - think about a sliding window (AI content window) over tons of data - we only want it to pick up what it needs to

(sliding window diagram text skipped but referenced)

- Brainstorming on types of data that we may want to collect & organize...

- general or global
- help files
- glossary
- tools
- developer's notebook
- ideas
- brainstorming
- documentation
- R & D (research and development)
- extended topics or deep dives
- chat messages
- news & updates
- history summaries
- emails
- tech support
- special tools
- business functions
- application players
- core concepts
- FAQ's
- tutorials
- step-by-steps
- Screenshots
- graphics & visuals
- sales marketing materials
- internal
- admin tools
- AI summaries
- user stories or scenarios

For me - add links, references, sources - document things at different levels

- Recruit help!

- Send an email out to our team - looking for (fill in the blanks)

- Send content & information to me & I'll enter it into the system (at first or for the first round)

- When we are ready, and if they (our team) wants, we can help train them on how to add, edit, & view the global knowledge base stuff

New notes from 12/12/25

- Small note or context - I was doing some scripture study and looking up some information about being alive & times & seasons - It took me to 2 Nephi 25:25 - then started to look at some verses both before and after that - Some of that was talking about scattered Israel (like in the Bible) I had this thought about adilas... and what I am working on. (see below)

- Things are all there... they are just scattered - we need to gather them & help them to be more organized - once they are gathered and organized, the goal is that the knowledge & information will continue to grow & thrive - help gather and organize!

- On training the AI agent I may have to tell it something & then tell it what I told it - light review & reinforce the key points

- I want to fix and fill in the gaps... That is awesome - At the beginning, anything will help - there aren't gaps, there are holes - just start

- We can always circle back around & refine, fix, patch, polish, and/or rebuild - like try storming, we plan, we do it, we see how it works, & we plan on circling back around again & again - small feedback cycle - build & break, build & break

- Terms (words) that I was looking up... alive or living... learning, growing, being used, purpose

- I do a lot of trail building and maintenance - it is quite the process... We have to scout things out, explore, make mistakes, flag, clip brush, do dirt work, build walls, bridges, crossings, bank retention etc. Then we need to use it (traffic), get others to use it, & then circle back around again - very similar to the try storming concept

- What does it take to become a master? What a process that is...

- Repetition & context... experience, learning & applying what you know

- Times & seasons - everything in its time

- Small side note... when I don't know how to spell something, I will enter it into Microsoft Word & have it help me with spelling - I already had the words reinforce and recruit. I couldn't remember how to spell Israel - Like gather Israel, from the Bible

- Anyway, I right clicked on the word Israel to see if it had any synonyms - It said Zion as one of the synonyms - I then had: Reinforce, Recruit, Israel (Zion) - Wow, I had a fun spiritual experience - it was awesome & it hit me - I don't know the full meaning, but that hit me pretty good - kinda like a mini outline of the plan of God

- Recruit help from others - old power users, reps, consultants, clients, trainers, etc. - If needed, reach out individually to see what you get back.

- I'm loving the deeper connection(s) like pondering inspiration & revelation - for me, help others tap into that source or channel

- Help create the rich, context specific environments, and make it possible for others to pull from the source

- We are gathering

- We are gathering, even in our choice of models -

(drawing of two models with text)

Other systems - mashups & using a bunch of different systems to run your business

Adilas.biz - systemizing the application & bringing business functions together

- We are organizing

- We are learning

- We are being led

- I am so grateful!!!

- Little by little, line upon line, here a little, there a little

- Like Daniel & King Nebuchadnezzar (had to lookup that spelling), king of Babylon (or the world) - In Daniel chapter 2 - Daniel had to both show (reveal, uncover, expose, disclose, tell, show) the dream & interpret (understand, explain, clarify, read, decode) what it meant - both reveal & interpret

- Stone soup - Bring what you have got & throw it in the pot - We are making stone soup - mixing & blending different pieces together over time

- I sure enjoyed my learning & connection session - It meant a lot to me - Hours of rich connection & inspiration - I love it!

-----------------------------------------

The above is the end of the notes. A few days later, I took the original notes over to ChatGPT, I plugged them in and we had a small chat about where we are headed. If you want, this is a follow-up discussion that we had on 12/19/25 - https://chatgpt.com/share/69462819-9640-8007-9f14-948f203349b9

Small change to the developer's notebook pages. Made it so that the page number gets passed along for pagination (if there are multiple pages per query). Made the change and pushed up new code for testing. Switched over to more planning and grabbing information off of post-it notes for my plan for training the AI agent or agents (plural). Great little session. Thoughts and inspiration were just coming to me. I loved it. See notes on element of time # 12527 for details.

Went over to Heather's parent's house. A construction company was fixing the road by our house. We didn't have any electricity. I brought my laptop, a bunch of old post-it notes, and a notebook. Spent some time going over my notes and transposing them from post-it notes to a physical notebook. I'm kinda glad to slow down for a bit. Rewriting the notes, it is helping me really internalize them, meaning my notes. I want this to work vs just a quick fix for the AI training.

Quick phone call with Josh to touch base on the training and where things are going.

Doing some planning on setting up a new database table to get and hold the AI global context. Brainstorming and light research.

Coding the help files to use the centralized global content database vs the individual databases per server for help files. Pushed up and merged in the new code.

Meeting with Russell and Abby. We started out and Abby did a show and tell on both vectors and photo editing (some of the things she is working on). Then Russell was showing her some cool links and some samples. Here are some notes from Russell:

- Keep practicing and challenge yourself.

- Take your challenge and then look-up how to fix it and how to improve.

- Russell was showing us how he gets inspired and how he goes in and grabs pictures, graphics, websites, animations, and then makes all kinds of collections and such.

- Fighting towards the goal.

- Helping people see your best work - when they see it, they say - okay, you're hired.

- Go find inspiration, energy, and emotions and then build or create it.

- Selling people off of the sizzle.

- Taking something that you like and molding it into something else.

- Keep getting better at things. Build lots of things.

- Inspiration helps you go beyond your own boundaries.

- Talking about styles and being consistent.

- There was some great advice for Abby and myself. Keep going... you are smashing it.

- Transitioning from design work into code work. Good path and good stuff.

Meeting with Greg from the Calder Group. We did some catching up, chatting, and shooting the bull. We then talked about some of the new happenings inside of adilas. Greg wanted to hear a little bit about our story... we had fun going back and forth. Greg did a bunch of accounting in school. We chatted about that and how things have changed and progressed over time. I also briefly talked about some history inside of adilas and how things grew and became to be. I sent him the link for the adilas lite (fracture) master plan. Good meeting.

Created a new help files table on a global context database. This should help with updates and making one standard place for AI training and content. Sent an email out to the guys who need to know.

Quick phone call with Steve. He is going to try to use Grok to do an AI agent. We are just experimenting. Working on new global context table inside of adilas to help with AI global context stuff.

Quick meeting with the Nxtlinq guys. They were showing us a new compare feature in the backend dashboard. That should help. They wanted some feedback but I have been working more with the main ChatGPT vs the mini private model. It takes time.

AI training inside of adilas. I was testing it, trying to see what it would retain and giving it some new information. I also logged into the backend and uploaded two PDF files. See chat session 9 (see attached).

On a GoToMeeting with Steve. He had a small question about some new settings. We did some debugging and got it working for him. He is working with a client that is using an ATM and then paying for invoices and giving change due back to the user. It's a special payment option from inside the cart. Basically, a new button, if turned on, that allows you see the invoice total and then a series of buttons in increments of $5 up to $250. The change due is calculated and the invoice is stamped with the correct information.

Steve and I also spent some time talking about AI training and using ChatGPT to help train the mini model.

Back working with ChatGPT on some training and PDF creation stuff. It is helping me create documentation for quick searches and special keywords and prompts. Working on inventory and operations functions like purchase orders, parts and general inventory, stock/units, recipe/builds, etc.

Working with Bryan to look at some future development for a school that tracks native american students and some of their goals and progress. Going over plans and talking about ways to setup the one-to-many relationships that they need. We briefly went over some document management (custom documents and assignments) and connecting to custom paperwork options. We also talked about stepping away from the computer to do some planning. Sometimes that really helps. Otherwise, you get distracted.

Working with Shannon. Showing her the progress with the AI agent training. Then jumping into new changes like the add bulk po line items. She wanted to see some of the new features and such. We then switched over and worked on more content for the adilas lite or fracture buildout. We are getting close. We only have one more major section on the fracture page to finish up our first-round planning for that page. Today we finished up the lessons learned, the life cycles and maintenance section, and started working on the white labeling section of the plan. See the fracture page in the adilas section for details.

Emailing the team about the new functions for bulk adding PO line items (mini reverse shopping cart functionality). Small email to Wayne about a possible global database for helping to train the AI agent. I would like his opinion and thoughts. Spent the rest of the time working on a training plan to help train the adilas AI agent. Spent time working with ChatGPT to make a plan and start rolling it out. Towards the end of the session, I was having ChatGPT help me generate some PDF's to help with some of the adilas quick search functions.

As a reference, here is the share link for the AI training session that I was having. Some good stuff there. Light elevator pitch, a few small PDF files, some help files, and making plans. https://chatgpt.com/share/69377fb0-4594-8007-81af-a48edbc72690

Emails and paying bills.