Meeting with the Nxtlinq guys. Going over frustrations with global context memory options... They will look at possible solutions. Ted was pitching some stuff... He wants us to focus on SOP's, ask our clients for their SOP's, flow charts, processes, documentation (good stuff), and helping to walk the system through the reasoning behind the decisions and the context.

We love the idea, we just need the AI agent to be able to really store all of the stuff that we are throwing at it. It can remember stuff, but it is super tiny. Also, because it is so tiny, and we have such a big application, it will load things into the global context window (its small memory space) and we will ask it about something else. It then kicks out the old to make room for the new. It really needs to be stored in small chunks in a database and then be able to get to that information on the fly. That's where we are heading. Frustrating but it will work out, it will just take time.