Research

Gradien studies how proprietary knowledge can improve AI work over time.

Explore research on context, workpaths, tacit knowledge, knowledge graphs, workflow learning, task preparation, and the infrastructure that helps AI build on prior work.

Research Proprietary knowledge emerges when context and workpaths remain connected. Companies keep feeding AI tools their explicit knowledge, documents and playbooks, while the tacit knowledge that actually drives decisions stays scattered in people's heads. This piece argues an AI system only becomes genuinely useful to one specific company when a knowledge graph and the workpaths that generate it stay connected, with humans approving what gets kept. 2026-07-20 · 7 min read

Research Compounding AI work requires a system that captures, understands, and learns. This piece argues that AI work only compounds when a system does more than retrieve stored information. It lays out the capture, understand, learn loop behind Core and explains why human approval, not automatic updating, is what turns AI output into trustworthy company knowledge. 2026-07-20 · 6 min read