layer: services · case study

Onsite OS

dev

2026

Role-based AI access control for small teams — gives employees a safe way to run AI on company files, managers a way to oversee the work, and admins a way to audit the full trail. The non-enterprise version of what big AI platforms sell top-down.

AI
Role-based access
Audit trail

context

Small businesses don't have a safe way to give employees AI access. Enterprise platforms sell access control top-down for big orgs — nothing exists for the 5-to-50-person team that wants employees running AI on real company files, with oversight. This is the gap.

approach

01Designed a three-role architecture to give small teams the same oversight enterprise platforms charge for — employees run AI on file-specific tasks, managers monitor the same work in real time, admins oversee the full operation with audit trails.
02Went deep on agent infrastructure trying to make it production-grade — built the backend on OpenClaw, Ollama, and local LLM routing to learn where orchestration frameworks actually break under load.
03Ditched OpenClaw when it couldn't hold up, and kept the lessons — where multi-agent systems fail, how to route between local and cloud models, how context engineering actually works when the context budget runs out.
04Rebuilding on a more practical stack. Pre-company, no customers yet — this is what I work on most.

outcome

Still in development. The product isn't live, but the agent infrastructure knowledge I gained trying to build it is what I bring to everything else now. The scar tissue is the asset.

All services