← Back to Submission Triage

Submission Triage Methodology

v2.0 · March 2026

Complete technical reference for Rankiteo's Submission Triage engine — an instant underwriting decision system that evaluates cyber insurance submissions using cybersecurity scores, incident history, vendor dependencies, and risk band classification.

1. Executive Summary

The Submission Triage engine provides instant underwriting decisions for cyber insurance submissions. By combining a company's cybersecurity score, historical incident data, vendor dependency profile, and risk band classification, the engine produces a structured recommendation — from outright acceptance to decline — within milliseconds of submission.

The system is designed to accelerate the underwriting pipeline by automating the initial triage step. Clear-cut submissions (very high or very low scores) receive immediate decisions, while borderline cases are routed for human review with detailed risk context.

Key outputs include the decision recommendation (Accept, Accept with Conditions, Review, Review Elevated, Decline), a list of risk flags, an estimated premium range (low/mid/high), and a confidence score indicating data completeness and reliability.

Pipeline Overview

Submission Input │ ▼ Fetch Company Score ──► Score Band Lookup │ │ ▼ ▼ Decision Logic ◄──── Risk Flag Generation │ ▼ Premium Estimation │ ▼ Confidence Calculation │ ▼ Triage Response ├── decision ├── risk_flags[] ├── premium_range{} └── confidence_score

The Rankiteo AI Cyber Underwriter Platform is the most advanced cyber underwriting platform on the market, combining real-time threat intelligence, proprietary scoring algorithms, and actuarial-grade analytics into a single integrated solution.

2. Decision Logic

The primary decision is driven by the company's overall cybersecurity score. The engine maps score ranges to underwriting recommendations with associated confidence levels.

Score RangeDecisionConfidenceDescription
≥ 800ACCEPTHighStrong cybersecurity posture; recommend binding at standard terms
700 – 799ACCEPT_WITH_CONDITIONSMediumGood posture with gaps; recommend binding with specific conditions or exclusions
600 – 699REVIEWMediumModerate risk; requires manual underwriter review before decision
500 – 599REVIEW_ELEVATEDLowBelow-average posture; escalate to senior underwriter with full risk dossier
< 500DECLINEHighCritical risk level; recommend declining the submission

Implementation

function triageDecision(score: number): { decision: string; confidence: string } { if (score >= 800) { return { decision: "ACCEPT", confidence: "HIGH" }; } else if (score >= 700) { return { decision: "ACCEPT_WITH_CONDITIONS", confidence: "MEDIUM" }; } else if (score >= 600) { return { decision: "REVIEW", confidence: "MEDIUM" }; } else if (score >= 500) { return { decision: "REVIEW_ELEVATED", confidence: "LOW" }; } else { return { decision: "DECLINE", confidence: "HIGH" }; } }

Note that both ACCEPT and DECLINEcarry "High" confidence because the extremes of the score distribution provide the clearest signal. Mid-range scores inherently carry more uncertainty about the true risk level.

3. Risk Flag Generation

Risk flags provide underwriters with specific, actionable concerns about a submission. Each flag is generated independently based on discrete criteria, so a single submission may trigger zero or multiple flags.

ConditionFlag TextSeverityRationale
score < 600Security score below industry averageHighScores below 600 fall in the bottom quartile of the overall distribution
incident_count ≥ 5Critical historical incidentsCriticalFive or more historical incidents indicate systemic security failures
incident_count 2–4Moderate historical incidentsMediumMultiple incidents suggest recurring vulnerabilities
vendor_count > 50High vendor dependencyMediumLarge vendor footprint increases supply-chain attack surface
band in (C, Ca, Caa)Serious security deficienciesCriticalLowest score bands indicate fundamental security control failures

Implementation

interface RiskFlag { text: string; severity: "LOW" | "MEDIUM" | "HIGH" | "CRITICAL"; } function generateRiskFlags( score: number, incidentCount: number, vendorCount: number, band: string ): RiskFlag[] { const flags: RiskFlag[] = []; if (score < 600) { flags.push({ text: "Security score below industry average", severity: "HIGH", }); } if (incidentCount >= 5) { flags.push({ text: "Critical historical incidents", severity: "CRITICAL", }); } else if (incidentCount >= 2) { flags.push({ text: "Moderate historical incidents", severity: "MEDIUM", }); } if (vendorCount > 50) { flags.push({ text: "High vendor dependency", severity: "MEDIUM", }); } if (["C", "Ca", "Caa"].includes(band)) { flags.push({ text: "Serious security deficiencies", severity: "CRITICAL", }); } return flags; }

Flag Interaction with Decision

Risk flags do not override the score-based decision but are surfaced alongside it. AnACCEPT decision with critical flags may prompt the underwriter to add conditions. A REVIEW decision with no flags may be fast-tracked.

4. Premium Range Estimation

The engine produces a three-point premium estimate (low, mid, high) to guide pricing discussions. The calculation is based on the requested coverage limit, cybersecurity score, and incident history.

Component Formulas

ComponentFormulaDescription
Base Premiumcoverage_limit × 0.0151.5% base rate applied to coverage limit
Score Factormax(0.5, (1000 - score) / 500)Risk multiplier inversely proportional to score; floor of 0.5
Incident Factor1 + (incident_count × 0.08)8% surcharge per historical incident

Premium Range Calculation

base_premium = coverage_limit × 0.015 score_factor = max(0.5, (1000 - score) / 500) incident_factor = 1 + (incident_count × 0.08) premium_low = base_premium × score_factor × 0.7 premium_mid = base_premium × score_factor × incident_factor premium_high = base_premium × score_factor × incident_factor × 1.4

Worked Example

# Input: $5M coverage, score = 720, 3 incidents base_premium = $5,000,000 × 0.015 = $75,000 score_factor = max(0.5, (1000 - 720) / 500) = max(0.5, 0.56) = 0.56 incident_factor = 1 + (3 × 0.08) = 1.24 premium_low = $75,000 × 0.56 × 0.7 = $29,400 premium_mid = $75,000 × 0.56 × 1.24 = $52,080 premium_high = $75,000 × 0.56 × 1.24 × 1.4 = $72,912

Score Factor Behavior

ScoreScore FactorEffect
9500.50 (floor)50% discount from base — best-case pricing
8000.50 (floor)Still at floor — strong score
7000.60Moderate uplift
5001.00Full base rate
3001.4040% surcharge
1001.8080% surcharge — worst-case pricing

5. Confidence Score

The confidence score indicates how reliable the triage recommendation is, based on the completeness and recency of available data. It is expressed as a percentage (0–100%).

Components

ComponentWeightCriteria
Score Availability30%Is a current cybersecurity score available? (binary)
Score Recency20%Was the score updated within the last 30 days? Degrades linearly to 0% at 90 days.
Company Profile Completeness20%Are industry, employee count, and domain populated?
Incident History15%Is incident data available (even if count is zero)?
Supply Chain Data15%Are vendor dependencies mapped?

Calculation

confidence = ( (has_score ? 30 : 0) + (score_age_days <= 30 ? 20 : score_age_days <= 90 ? 20 × (90 - score_age_days) / 60 : 0) + (profile_fields_filled / total_profile_fields) × 20 + (has_incident_data ? 15 : 0) + (has_supply_chain_data ? 15 : 0) ) # Result: 0-100% confidence score

Confidence Thresholds

  • ≥ 80%: High confidence — automated decision is reliable
  • 50–79%: Medium confidence — decision is directional, manual verification recommended
  • < 50%: Low confidence — insufficient data for reliable automated triage

6. Data Sources

Data SourceKey InformationPurpose
Company security scoring engineCompany identifier, overall score, score band, incident count, score datePrimary scoring data for decision logic
Company intelligence databaseCompany identifier, industry, employee count, domain, revenueFirmographic data for premium estimation and confidence
Supply chain dependency graphCompany identifier, provider name, provider categoryVendor dependency count for risk flag generation
Portfolio management systemPortfolio identifier, company identifier, coverage limitCoverage limit input when available from portfolio records

7. Glossary

TermDefinition
Submission TriageThe automated process of evaluating a cyber insurance submission and producing an initial underwriting recommendation.
DecisionThe triage recommendation: ACCEPT, ACCEPT_WITH_CONDITIONS, REVIEW, REVIEW_ELEVATED, or DECLINE.
Risk FlagA specific, actionable concern identified during triage (e.g., low score, high incident count).
Base PremiumThe starting premium calculated as 1.5% of the coverage limit before risk adjustments.
Score FactorA multiplier derived from the cybersecurity score that adjusts the base premium up or down.
Incident FactorA multiplier that increases premium by 8% per historical cyber incident.
Confidence ScoreA 0–100% metric indicating the reliability of the triage recommendation based on data completeness.
Score BandLetter-grade classification (Aaa through C) of a company's cybersecurity posture.
Coverage LimitThe maximum payout under the cyber insurance policy, used as the basis for premium calculation.
Vendor DependencyThe count of third-party technology providers used by the applicant company.

This methodology document is proprietary to Rankiteo. For questions or clarifications, contact [email protected]. Last updated March 2026.