Methodology
Last updated: July 22, 2026
ADAME is the framework Praevum uses to turn open-source reporting into a structured, defensible read on how a situation is moving. It scores five dimensions of a situation, combines them into a single composite, and classifies that composite into a plain-language risk band. This page explains the framework in concept. It does not publish the internal weights, anchors, or formulas that produce a specific number.
The five pillars
Every assessment is built from five pillars. Each measures a distinct dimension of a situation.
Aggression
the intensity of hostile or coercive action by the actors involved.
Destabilization
the degree to which existing order, institutions, or control are being eroded.
Alliance Pressure
the strain placed on alliances and partnerships, and the risk of realignment.
Momentum
the speed and direction of change, whether a situation is escalating or settling.
Evidence
the reliability, corroboration, and timeliness of the underlying reporting.
The composite score and its bands
The five pillars resolve into a single composite on a 1.0 to 5.0 scale. Each composite falls into one of five bands.
- Stable (1.0 to 1.5) — low risk, routine monitoring.
- Elevated (1.6 to 2.5) — some risk, prepare contingencies.
- Concerning (2.6 to 3.4) — moderate risk, advise contingency actions.
- Critical (3.5 to 4.4) — high risk, warranting strategic intervention.
- Crisis (4.5 and above) — imminent or ongoing, requiring immediate response.
Evidence measures reliability, not severity
The Evidence pillar is different from the other four. It does not describe how dangerous a situation is. It describes how much confidence the underlying reporting can bear — whether claims are corroborated, how current the information is, and how trustworthy the sources are. Strong evidence does not raise a score, and weak evidence does not lower it. Evidence governs how much weight the rest of the assessment can reasonably carry.
A deterministic engine, with AI as the analyst layer
Every ADAME score is produced by a deterministic scoring engine. The same inputs always yield the same score. Artificial intelligence serves as the analyst layer around that engine — gathering reporting, extracting indicators, drafting narrative, and surfacing what matters. The AI never sets the score. Scoring is the engine's responsibility alone, which is what makes a Praevum assessment reproducible and auditable.
Alignment with established analytic practice
The framework follows the habits of disciplined intelligence work — expressing uncertainty explicitly, grading sources for reliability, and separating what is known from what is inferred. Praevum applies these practices as a matter of method. It does not hold, and does not claim, any external certification or accreditation.