Where AI belongs, where it doesn't, and how to tell the difference.
Organizations are adopting AI while still figuring out what it means for organizational design and where AI actually fits. I keep returning to the same underlying problems: process design, conceptual modeling, semantic ambiguity, and where humans fit into systems that increasingly make interpretive decisions on their behalf. The point is not formalism for its own sake, but to clarify my own thinking about how data can be made precise enough, and decision paths explicit enough, for environments where correctness, auditability, and accountability matter.
There is another problem that I think is being overlooked: understanding what an organization actually is at its essential level, the network of actors, acts, and commitments that would persist even if every piece of software supporting it were replaced. Redesigning for or with AI without that understanding tends to produce systems that automate a step faithfully while missing what the organization needed in the first place.
This site is a personal effort to work through that part of the problem: how to think more clearly about semantic precision, testable constraints, and human accountability in AI-supported systems.
The synthesis draws on two traditions that rarely talk to each other: philosophy of meaning, and formal data, conceptual, and enterprise modeling. See my key influences.
A term can mean different things in Finance and Operations, and neither team is wrong; meaning follows practice, not a fixed definition.
Under the closed world assumption, an unrecorded fact is false, not merely unknown. When money, rights, or compliance are at stake, "I don't know" isn't a good enough answer.
Automating a step doesn't remove the commitment behind it; it only changes who, or what, performs it. Capable automation still needs a named party accountable for the outcome.
A schema's declared meaning and its actual use can drift apart over years. Usage and operational traces often show that drift most clearly, though closing the gap still means making the definition explicit again.
These essays and notes explore organizational design, AI governance, semantic modeling, and the practical challenge of making enterprise data clear enough to support accountable systems. The clearest path through the core argument is the two-part series below, which moves from semantic foundations to the automation boundary.
I draw on two traditions that rarely talk to each other: philosophy of meaning, and formal data, conceptual, and enterprise modeling. None of these thinkers were solving the problem I'm working on; this site is my attempt to adapt their ideas to enterprise AI and data systems.
If you are working through similar questions, especially around data platform design, semantic modeling, AI governance, or accountable automation, feel free to send me a note. This site is a personal side project, so I may not always reply quickly, but I am glad to hear from people exploring related ideas.
If any of the influences above are ones you keep coming back to as well, mention that in your note. I like hearing from people converging on similar territory from different directions.
Email: gsawatzky@knowledge-foundation.ai
I also keep a private Collaboration Zone with notes and essays shared only with people I'm actively working with. It's invite-only: sign in there with Google to request access, or just email me directly and I'll set you up.