Inferal

Inferal builds what it calls a data-native operating system for AI agents. The platform connects to a company's existing databases and message queues and maintains live operational models of the entities, relationships, policies and business rules those systems describe, so agents act on current data rather than on stale snapshots pulled through request-response queries. Rather than having agents poll, Inferal continuously matches declared rules against incoming data and activates the relevant agent with its context attached when conditions align. Every activation and decision is recorded against an ontology-based semantic layer so the reasoning is traceable end to end. The company was founded in 2023 by Yurii Rashkovskii and grew out of Omnigres, his earlier open-source Postgres-extension project; as of mid-2026 it is pre-general-availability and engaging customers through a design partner program.

Inverts the usual data access pattern for agents: instead of an agent querying a database and waiting, business rules are declared once against an ontology and continuously evaluated, so the system wakes the agent at the moment conditions are met and hands it the context. Combined with implication graphs and full audit trails, that makes each decision explainable, which is the gating requirement in the regulated domains it targets.