By default, most AI conversations start from zero. Some tools (e.g., Claude) offer memory features or can read from a connected folder, but they work with what you choose to share in that moment, not the full depth of knowledge your business has built up over time.
A different approach is starting to emerge. Tools like Obsidian (note-taking application), combined with local AI models running on your own machine, make it possible to build a knowledge base your AI can actually query. Not a chat session, but a persistent store of structured notes: budgets, supplier research, process documentation, meeting outcomes, decisions made and why. When you ask a question, the AI searches that store and answers from your own accumulated knowledge before reaching for anything else.
The practical difference is real. When we were scoping the PII scrubbing functionality for Pixel Assist, we described what we were building to our vault-connected AI. It flagged that we had tackled something similar for a previous client and asked if we wanted to reuse the research from that project: the libraries that had worked best, the data flow requirements we had already worked through, the lessons from that engagement. Work that would have taken days to reconstruct was available in minutes. Your data also stays on your machine. Nothing is sent to a third-party server, which matters when the notes contain client information or sensitive business thinking.
The setup is not simple. Obsidian handles the notes, Ollama handles the local model, and Hermes (AI Agent) connects them. Getting all three working together takes some patience, and choosing the right model for your machine is not obvious. But the direction is clear: this kind of capability will become easier to package and deploy. When it does, the question worth asking is whether your knowledge is in good enough shape to be worth querying. That is a problem worth solving now, regardless of the tooling. In "open knowledge format: a format for company knowledge that your team and your AI can both use" we discuss why the Open Knowledge Format (OKF) is a good place to start.
Still have questions? Give us a call, , or send us a message to understand more without any pushy sales tactics.
Whether you've got a detailed brief or just a rough idea, we're happy to have a conversation.
No pitch, no pressure, just a straight talk about what you're working on and whether we're the right fit.