Living Models
Causal Intelligence for the Decisions That Actually Matter
Most analytics systems are built to describe the past. Living Models is built to reason about what happens when you change something.
Causal Decision Architecture
Living Models is a causal decision-support architecture that is causal, counterfactual, continually updated, and treatment-oriented.
CAUSAL, NOT CORRELATIONAL
Living Models encodes mechanisms, not correlations. Recommendations are expressed as estimated interventional effects P(Y | do(X)) — not patterns observed in historical data.
Learn about the causal foundation →THE KNOWLEDGE ACQUISITION TOOL
A structured expert interview that extracts implicit causal knowledge and converts it into a first-draft DAG in approximately 45 minutes — without requiring the expert to learn what a CPDAG is.
Explore the interview architecture →COUNTERFACTUAL REASONING
Living Models can reason about what would have happened under conditions that never occurred — evaluating the cost of decisions not taken as rigorously as decisions that were.
See how counterfactuals work →TREATMENT-ORIENTED OUTPUT
The output is not a description or a prediction. It is a ranked list of interventions evaluated by Expected Value of Intervention, compared against the counterfactual trajectory.
Understand intervention ranking →Who Benefits from Living Models?
Living Models creates value for everyone involved in the decision-making ecosystem.
FOR STRATEGY EXECUTIVES
- —Get causal answers to intervention questions, not correlation summaries
- —Understand which levers actually move the outcomes you care about
- —Receive ranked recommendations with auditable evidence and assumptions
- —Know the expected cost of inaction, not just the expected value of action
FOR DATA SCIENTISTS
- —Start from an expert-originated causal graph instead of building one from scratch
- —Use the Knowledge Acquisition Tool output as structured input to NOTEARS / PC / GES
- —Apply Double Machine Learning on a correctly specified causal structure
- —Move from author to editor of the causal model
FOR ORGANIZATIONS
- —Invert the standard workflow: extract knowledge first, formalize second
- —Prevent wrong-model confidence — precise wrongness with tight confidence intervals
- —Build decision intelligence that updates as new data arrives
- —Surface the boundary of what your analytics can and cannot see
The Question Was Always the Problem
The data was never the problem. It was always the question. Living Models is the attempt to build decision-support architecture that tells you not just what happened and what is likely to happen — but what you should do about it, and why that action and not another.