Retrieval Is Not Enough
Why Organizational AI Needs Epistemic Infrastructure
Paper, preprint — Federico Bottino, Carlo Ferrero, Nicholas Dosio, Pierfrancesco Beneventano — arXiv preprint (cs.AI), 2026
Organizational AI systems need epistemic structure beyond retrieval capabilities. We present OIDA, a framework that represents organizational knowledge as typed objects with epistemic properties, commitment strength, and contradiction status. A key innovation is QUESTION-as-modeled-ignorance, which surfaces organizational unknowns. Experimental evaluation uses an Epistemic Quality Score methodology; OIDA achieves 0.530 EQS with 3,868 tokens versus 0.848 for a full-context baseline using 108,687 tokens. The QUESTION mechanism shows statistical significance with Fisher p=0.0325.
Epistemic infrastructure · Organizational AI
arXiv ↗ · PDF ↗ · Read more →
