How Sophurion Horizon thinks.

Horizon scores risk on a 0–100 scale across five Predictive Sustainability Intelligence (PSI) domains. Each domain receives an Exposure score (how concentrated the underlying risk factors are for this organization) and a Preparedness score (the strength of controls, plans, and capabilities offsetting that exposure).

Each intelligence domain is evaluated using a proprietary scoring framework that balances risk exposure against organizational preparedness. Exposure factors increase risk, while preparedness measures reduce risk. Domain scores are normalized to a 0–100 scale and combined using industry-specific weighting models.

The Overall Risk Score is generated through a weighted aggregation of all five intelligence domains. Weighting varies by industry and organizational profile to reflect sector-specific risk realities.

The Emerging Risk Index is a forward-weighted blend biased toward Climate, Geopolitical, and Regulatory signals — the three domains where trajectory typically diverges most from current state. It surfaces issues that current-state scores understate.

The Financial Exposure Estimate is an order-of-magnitude figure derived from declared revenue band, industry-baseline disruption coefficients, and the overall risk score. It is a planning anchor, not an actuarial loss estimate.

Every dashboard carries a 0–100 Confidence Rating. It tells you how much weight to put on the numbers — not how worried to be. Three inputs feed it:

Confidence Ratings are derived from assessment completeness, geographic specificity, and organizational context. Higher-quality inputs produce higher-confidence outputs and more reliable recommendations.

Bands: 80–100 High — decision-grade. 60–79 Moderate — directionally reliable; verify headline calls. < 60 Indicative — treat as a draft; complete the assessment before acting.

Recommendations are generated by a deterministic rules engine, then enriched with an AI-written executive narrative. The deterministic layer keeps the same inputs producing the same actions — a property we consider non-negotiable for decision support.

The engine sorts every action into four categories and four priority levels:

Priority is assigned from the underlying domain risk score, with explicit escalation when the score crosses pre-set thresholds:

Every score generated by Horizon can be traced back to the underlying assessment inputs that influenced it. Organizations can review the drivers behind individual risk ratings, recommendations, and confidence levels, providing transparency without requiring specialized data science expertise.

The five PSI domains are not silos — they are lenses on a single converging system. Every assessment signal is mapped to at least one domain; many feed multiple.

As the Sophurion learning engine aggregates anonymized organization data, domain weights and benchmark distributions are re-fit by sector. Methodology versions are tracked; reports include the model version used to generate them.

Horizon is an executive intelligence platform in early launch, designed to support organizational decision-making while the platform continues to expand its intelligence capabilities. It is not insurance, not an audit opinion, and not a substitute for jurisdiction-specific legal or financial advice.

• Scores reflect self-reported inputs. Garbage in, garbage out — the Confidence Rating exists to make that visible.
• Industry baselines are calibrated from a growing dataset; sectors with thin coverage default to broader peer groups.
• The Financial Exposure Estimate is an order-of-magnitude planning anchor. Treat it as a band, not a forecast.
• AI-generated narrative summarizes the deterministic model output. The underlying numbers, not the prose, are the source of truth.

The methodology is co-developed with Pearce Sustainability Consulting Group and is reviewed continuously as the platform matures.