From Data to Actionable Predictions
Flexible AI models to solve business-specific challenges across the insurance lifecycle.
Overview
Pinpoint’s Discovery Models let carriers define and predict any two-outcome event, unlocking use cases from churn and premium leakage to risk control and beyond. With Binary Classifiers at the core, teams can experiment, validate, and deploy predictive solutions quickly and safely.
Whether in underwriting, claims, or retention, Discovery Models empower insurers to innovate, align decisions with data, and accelerate value creation.
Quickly Move from Discovery to Production
This approach unlocks the potential to innovate across underwriting, claims, and retention workflows.
- Binary Classifiers
Custom Predictive Modeling Software
Pinpoint’s Binary Classifiers provide a software-first platform for custom binary classification modeling, for carriers to develop business-specific predictive models for any defined two-outcome event, such as “Did Cancel” vs. “Did Not” or “Adopted Program” vs. “Declined.” This solution turns labeled outcomes into actionable insights, letting teams experiment, validate, and deploy binary outcome analytics without heavy development or relying on data science resources.
- Build and validate models for any binary outcome
- Rapidly move from discovery to production
- Test predictive ideas without custom development
Key Discovery Benefits
Rapid Experimentation
Test and validate predictive ideas quickly and safely.
Flexible Innovation
Build custom models for any two-outcome event.
Scalable Deployment
Move from discovery to production without heavy development.
Case Studies
Predictive Power at Work
Auto Case Study: Loss Cost Model to Optimize Marketing Spend
Pinpoint helped reduce Loss Ratio by 4 points on personal auto
Client: National Auto Carrier
Line of Business: Auto Insurance
Number of Policies: 714,577
Average Premium: $2,495
Use Case Focus: Customer Acquisition
Auto Case Study: Claims Severity Predictions
Pinpoint helped reduce Loss Ratio by 3 points with Claims Severity Predictions
Client: National Auto Carrier
Line of Business: Auto Insurance
Number of Policies: 714,577
Average Premium: $2,495
Auto Case Study: Litigation Risk Scores
Pinpoint Helped Identify an Additional 6% of Claimants Likely to Seek Attorney Representation
Client: National Auto Insurance Insurer
Line of Business: Auto Insurance
Model Trained: 130,000 Claimants
Home Case Study: Claims Frequency Prediction to Optimize Growth
Pinpoint helped reduce Loss Ratio by 3 points to optimize profitable growth
Client: National Home Insurance Carrier
Line of Business: Home Insurance
Number of Policies: ~300,000
Average Individual Premium: $2,900
Home Case Study: Loss Cost Prediction to Reduce Non-Weather Water Losses
Pinpoint helped reduce Loss Ratio by 4 points during pre-application
Client: National Home Insurance Carrier
Line of Business: Home Insurance
Number of Policies: ~150,000
Average Individual Premium: $1,291
Auto Case Study: Loss Cost Prediction to Optimize Growth
Pinpoint helped reduce Loss Ratio by 4.6 points by focusing on Individual Drivers
Client: National Auto Carrier
Line of Business: Auto Insurance
Number of Policies: 80,000
Average Premium: $1,500
Home Case Study: Claims Frequency Predictions for Book Acquisition
Pinpoint helped reduce Loss Ratio by 7 points by focusing on Book Acquisition Strategy
Client: National Home Insurance Carrier
Line of Business: Home Insurance
Number of Policies: 200,000
Average Premium: $2,813
Frequently Asked Questions
What are Pinpoint’s Discovery Models?
Discovery Models are flexible, actuary-built predictive tools that let carriers design and test custom models for any defined two-outcome event (churn vs. retention, program adoption vs. decline). They provide a safe, fast way to experiment with predictive solutions across underwriting, claims, marketing, and retention.
How do Binary Classifiers work?
When carriers can define an outcome as “did” vs. “did not,” Pinpoint can model it as a Binary Classifier. Upload labeled records and OnPoint trains and validates a custom classifier, returning results within 24 hours. With a few clicks, the model is production-ready via configurable APIs—turning hypotheses into real-world value fast.
What business challenges can Discovery Models help solve?
Pinpoint builds custom models around the outcomes your organization is prioritizing, aligning each build to your most consequential challenges. OnPoint trains on your labeled outcomes and promotes models to production rapidly, with no fixed catalog or constraints on what you can model. The result is a production-ready model precisely focused on your current priorities.
Do carriers need their own data science team to use these models?
No. OnPoint automates model build and validation end to end—feature prep, training, testing, and promotion—so both business and technical teams can experiment and deploy without custom development. You provide a correctly pulled labeled dataset (minimal internal effort), the platform returns results within 24 hours, and production enablement is a few clicks via configurable APIs.
How can carriers access Discovery Models?
Commercial clients receive annual allocations to build and run Binary Classifiers. Prospective clients can also use Discovery Models in Proof of Value engagements when paired with core Pinpoint solutions like Loss Predictions, Early Cancellation Risk, or Attorney Rep Propensity.