# 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 [Learn More](/content/case-study/auto-case-study-loss-cost-model-to-optimize-marketing-spend/index.html) ### 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 [Learn More](/content/case-study/auto-case-study-severity-predictions/index.html) ### 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 [Learn More](/content/case-study/case-study-likelihood-of-third-party-claimant-to-seek-attorney-representation/index.html) ### 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 [Learn More](/content/case-study/case-study-claims-frequency-loss-prediction-for-home-carrier/index.html) ### 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 [Learn More](/content/case-study/case-study-loss-cost-prediction-for-a-home-carrier/index.html) ### 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 [Learn More](/content/case-study/case-study-auto-loss-cost/index.html) ### 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 [Learn More](/content/case-study/case-study-claims-frequency/index.html) ## 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.