Sanjay Sabnani › AI research
AI Technology · Patent Pending

Causal Wisdom Harvester

Causal Neuro-Symbolic AI

The Causal Wisdom Harvester is a patent-pending system that produces structured causal models from domain texts — enabling AI systems to apply structured, traceable reasoning in constrained domains rather than probabilistic inference alone. This is the core Causal Neuro-Symbolic AI (CausalNeSy AI) work.

Applied to the International Regulations for Preventing Collisions at Sea (COLREGs), the Harvester produced a queryable causal graph from a single regulatory document: 1,804 entities, 488 causal rules, 85 behavioral patterns, and 2,318 relations. Given any vessel situation, the system traverses applicable rules, surfaces conflicts, traces consequences, and identifies negative outcomes — with full source attribution. The same engine has been applied to US patent law, medical literature, and contract corpora.

Persona Modeling

A variant of the Harvester processes interview and behavioral data into a deployable model of a person’s decision architecture — encoding values, cognitive patterns, beliefs, and interaction preferences as a causal graph an AI can reason from, not just retrieve against.

The “Sanjay” AI Twin chat on this site is powered by exactly this: a model built from interview data using this system. What you’re talking to is a working demonstration of the technology.

The extraction methodology is proprietary and patent-pending. Research collaboration and domain inquiries welcome.

Talk to the AI guide Contact

More by Sanjay Sabnani