01 — Problem
Pricing expected loss is not enough.
Insurers price the loss they expect. Prevention exists, but generic, hard to personalise, and disconnected from loss ratio.
AI-powered synthetic twins for insurance
From Risk to Prevention
Ensembling builds Twins — a B2B white-label platform that turns insurer data into dynamic synthetic twins, prevention scenarios, and measurable portfolio decisions.
Risk → Prevention
Live cohort · 18-month horizon
Live
Baseline (no action)
With Twins prevention
Loss reduction
−18.4%
Avoided loss
€5.8M
Confidence
91%
Product
Prevention, made measurable.
For each policyholder, Twins simulates two probabilistic risk curves over time: the baseline trajectory, and the trajectory under recommended prevention. The gap between them, summed across the portfolio, is your savings.
Motor liability
Baseline
With prevention
Avoided loss · 18 months
€5.8M
Modeled across 124,580 policyholders, ranked by intervention readiness.
Loss ratio Δ−18.4%
Combined ratio Δ−6.2pt
Interventions14,820
Model confidence
91%
Calibration (Brier)0.062
Coverage @ 95% CI94.1%
Backtested cohorts38
02 — Solution
A probabilistic twin per policyholder.
Twins simulates risk evolution and prevention scenarios, ranks interventions by avoided loss, confidence and readiness.
03 — Value
Prevention as a financial lever.
Identify rising-risk cohorts earlier, rank actions by expected impact, make prevention economically accountable.
Platform intelligence
A live operating system for synthetic twin decisions.
Twins brings portfolio KPIs, executive signals and prevention priorities into one operating view, so insurance teams can move from analysis to action without losing portfolio context.
Twins prevention cockpit · portfolio overview
A white-label intelligence layer for insurer-owned data, dashboards and customer channels.
Live intelligence
Policyholders monitored
124,580
+6.8% vs last month
Active synthetic twins
121,940
+4.2% twin coverage
Preventable claims identified
3,420
+11.4% opportunities
Estimated loss reduction
€5.8M
+0.7M modeled upside
Underwriting cases supported
486
+18 priority cases
Portfolio risk drift
2.1 pts
−0.4 pts after prevention
Executive narrative
Where Twins is changing portfolio economics this week
The platform identified a concentrated preventable-loss corridor in health and SME segments, with high-confidence interventions already reducing projected loss and stabilizing portfolio drift.
€4.2k
Health deterioration corridor
Clinical outreach is reducing short-term claims propensity in the acute health segment.
−1.1 pts
SME segment drift contained
Employer-led prevention bends drift before losses fully materialize.
486
Underwriting quality uplift
Synthetic twins support review decisions with cluster-level explainability.
Prevention center
From risk signal to owner-ready intervention.
The product is not just a dashboard: it converts synthetic twin signals into ranked actions, expected avoided loss, operational owners and confidence levels.
01
Secure insurer data
Connect within the insurer-owned perimeter and keep the customer relationship white-label.
02
Generate dynamic twins
Build probabilistic synthetic twins for policyholders, cohorts and portfolio segments.
03
Simulate scenarios
Compare baseline loss curves with prevention, pricing or monitoring changes.
04
Activate outputs
Push explainable next-best-actions into actuarial, underwriting and customer journeys.
Clinical outreach bundle
High-urgency intervention for acute health deterioration cases with clinical owner assignment.
94
Employer wellbeing sequence
Portfolio-scalable intervention targeting SME workforce engagement and care activation.
74
Driver behavior incentive
Low-cost telematics-based incentive to lower harsh events and night-routing risk.
76
Portfolio & underwriting
Explain portfolio pressure before claims make it obvious.
Segment heatmaps, geographic signals and underwriting cases translate twin-level dynamics into decisions for portfolio strategy, renewal committees and prevention operations.
Where the health portfolio needs attention
Health-only cohort pressure, profitability and prevention priorities.
30–44 · HealthStress-linked prevention opportunity
45–59 · HealthFamily health severity watch
60+ · HealthChronic-care escalation risk
Corporate HealthWorkforce absence and claims pressure
Corporate Benefits
€18.8M premium · 14% profitabilityExecutive Health
€14.6M premium · 18% profitabilityFamily Health
€11.2M premium · 9% profitabilityUnderwriting cases supported
Explainable synthetic-twin recommendations at case level.
Luca Bellini · Health
Renewal review: elevated claims trend · Recommended action: prevention pathway
61loss ratio
Irene Fontana · Health
Care plan adjustment · Recommended action: request updated clinical data
74loss ratio
Paolo Conti · Health
Family health expansion · Recommended action: Escalate review
82loss ratio
Sofia Greco · Corporate benefits
SME plan renewal · Recommended class: Class B
58loss ratio
Simulation & reports
Scenario modeling that becomes an executive readout.
Compare baseline and simulated outcomes, then package the insight into portfolio, underwriting and operations reports that stakeholders can act on.
Clinical prevention scenario
Applying targeted outreach to high-risk health cohorts.
Claim probability
28.4% → 22.1%
Expected loss
€1.28M → €1.04M
Preventive effect
11.2% → 16.8%
Portfolio drift
2.1 pts → 1.3 pts
- Applying clinical prevention reduces modeled claim probability by 6.3 points without immediate product redesign.
- Avoided loss is concentrated in Family Health and Executive Health, where intervention responsiveness is strongest.
- Uncertainty narrows as monitoring intensity rises, improving explainability for underwriting follow-up.
Team
A focused team for regulated AI products.
Ensembling is built by a small team across data science, software engineering, go-to-market and legal operations, with experience in research, regulated markets and product execution.