Use Cases

Insurance

Policy Renewal and Cancellation Risk Intelligence

Model6 can identify policies with an elevated likelihood of customer cancellation, lapse, or non-renewal and estimate renewal likelihood within a defined period. The model can provide priority bands and approved reason indicators for eligible policies.

Agent6, Connect6, and PolicyBuddy can connect validated signals to approved service, billing, agent, or communication workflows. The insurer retains control over price, terms, underwriting eligibility, renewal, cancellation, non-renewal, offers, contact permissions, and customer treatment.

Why This Use Case Matters

Attrition signals often appear before the policy outcome

Billing behavior, service interactions, price change, claims, engagement, channel activity, and prior renewal behavior can indicate elevated risk before the customer leaves.

Broad retention activity spends effort where it may not help

Treating every policy the same creates contact fatigue, workload, and concession pressure. Teams need to separate risk, eligibility, actionability, and economic value.

Retention requires controlled measurement, not correlation alone

A policy may renew naturally after outreach. Holdouts, action tracking, cost, margin, customer impact, and governance are needed to understand incremental value.

Retention Activity Is Often Late, Broad, and Difficult to Measure

Cancellation, lapse, non-renewal, product migration, and renewal are related but distinct outcomes. Each can require a different label, prediction window, intervention, owner, and safeguard.

Policy, billing, payment, claims, service, distribution, engagement, and outcome data sit across separate systems.
Outcome definitions vary by product, channel, reason, and jurisdiction.
Teams may receive a signal only after a missed payment, complaint, notice, or cancellation request.
High risk does not always mean an intervention is permitted, appropriate, or economically useful.
Scores can be hard to interpret without reason indicators and policy context.
Leaders need to separate natural renewal from incremental retention created by an intervention.

Bay6 separates the target outcomes and designs each model around a defined population, time horizon, decision point, and permitted action. Model6 produces validated cancellation-risk or renewal-likelihood scores with approved reason indicators and monitoring.

Agent6 can create tasks, queues, service actions, or feedback capture. Connect6 and PolicyBuddy can support approved policyholder interactions and authenticated service. Forge6 supports data readiness, intervention design, fairness review, workflow mapping, pilot measurement, and governance.

Operating principle: Model6 identifies and prioritizes retention risk. The insurer decides whether, when, and how to contact the customer or change any policy, price, term, or offer.

How the Renewal and Cancellation Risk Workflow Works

The workflow separates the policy outcome from the intervention and measures both independently.

Define policy outcomes

Set cancellation, lapse, non-renewal, renewal, eligible policies, time horizons, exclusions, contact rules, and permitted actions.

Prepare lifecycle data

Map and review approved policy, billing, claims, service, distribution, engagement, and historical outcome data.

Build and validate models

Develop outcome-specific models and test calibration, lift, stability, leakage, segment performance, and proxy risk.

Score and explain

Provide probability, priority band, and approved reason indicators through batch or API delivery.

Activate approved responses

Create agent, service, billing, product, or communication tasks using insurer-defined permissions and rules.

Measure impact and model health

Track holdouts, contacts, responses, cost, margin, outcomes, overrides, customer impact, drift, and versions.

What Model6 Helps Retention Teams Predict and Prioritize

The model separates risk from actionability so business teams can focus on permitted, relevant, measurable responses.

Cancellation risk prediction

Estimate the likelihood that an eligible policy may cancel or churn within the insurer-defined period.

Renewal likelihood prediction

Estimate the likelihood that an eligible policy will renew where data, outcome definition, and validation support the model.

Outcome-specific segmentation

Separate cancellation, lapse, non-renewal, renewal, product migration, and other approved outcomes.

Reason and actionability indicators

Provide approved factors that help teams understand the signal and identify an appropriate response.

Retention workflow activation

Create approved tasks, queues, conversations, reminders, service paths, or offer-review processes.

Intervention and holdout measurement

Link treatments, controls, costs, responses, policy outcomes, overrides, and model versions to estimate incremental value.

What Insurers Can Improve

Measure model quality, operating adoption, incremental retention, economics, and customer impact together.

Earlier cancellation, lapse, non-renewal, or renewal-risk visibility
Retention queue and outreach efficiency
Reason usefulness and percentage of risk with an approved action
Contact, response, service completion, and agent follow-up
Incremental retention, cost, margin, and lifetime-value impact
Customer contact, fairness, complaint, drift, and audit outcomes

Outcome focus: A governed retention workflow identifies actionable risk, directs approved service or agent activity, and measures whether the intervention created real policy value.

Pilot Measurement Areas

Use one product, outcome, prediction window, policy population, channel, intervention, holdout design, and operating owner.

Measurement Area	What to Track
Model performance	Discrimination, calibration, lift, precision, recall, stability, and performance by product, channel, reason, and segment.
Actionability	High-risk policies with an approved response, reason usefulness, missing context, professional agreement, and overrides.
Intervention performance	Contact, response, completion, timing, channel, agent activity, cost, and service outcome by treatment and holdout.
Policy outcomes	Renewal, cancellation, lapse, non-renewal, product movement, timing, and reason after the intervention.
Economic value	Incremental retention, premium, margin, cost to retain, concessions, lifetime value, and portfolio impact.
Governance and customer impact	Contact frequency, complaints, opt-outs, differential treatment, drift, incidents, and audit records.

Frequently Asked Questions

Have questions? Browse our FAQs to learn more about how Bay6 AI works and how it supports your organization’s goals.

What is policy renewal and cancellation risk intelligence?

It uses predictive models to estimate the likelihood of a defined policy outcome and connect eligible risk signals to approved retention workflows and measurement.

Are cancellation and renewal the same model?

No. Cancellation, lapse, non-renewal, renewal, and product movement should be defined separately with their own populations, time horizons, data, actions, and validation.

Can Model6 change a price, term, or policy status?

No. Scores support prioritization. Pricing, terms, eligibility, renewal, cancellation, non-renewal, offers, and policy actions remain with authorized insurer teams and systems.

How should intervention impact be measured?

Renewal after contact does not prove causation. Use approved holdout, test-and-control, or related measurement methods and track cost, margin, customer impact, and policy outcomes.

What should an initial pilot include?

Select one product, outcome, prediction window, population, channel, intervention, user group, holdout design, professional owner, integration, and governance plan.

Responsible AI and Policy Oversight

Model6 supports retention prioritization. Pricing, underwriting, renewal eligibility, terms, cancellation, non-renewal, offers, and policy authority remain with the insurer.

Define cancellation, lapse, non-renewal, renewal, and other outcomes separately with clear populations, horizons, exclusions, and permitted uses.
Use relevant, permitted, documented data and review sensitive attributes, proxies, contact constraints, and adverse impact.
Validate calibration, lift, stability, reason quality, actionability, segment performance, and intervention capacity before deployment.
Keep the score separate from price, terms, underwriting eligibility, renewal, cancellation, non-renewal, and policy-status decisions.
Apply approved contact permissions, frequency, channel, disclosure, accessibility, complaint, and customer-treatment rules.
Monitor scores, reasons, actions, holdouts, economics, overrides, customer impact, drift, incidents, and complaints by version.

Explore More AI Solutions for Insurance

Extend the workflow through counselor task creation, follow-up coordination, system updates, document requests, and enrollment operations handoffs.

EXPLORE agent6

Supports conversational self-service and knowledge-based student support where broader student services or campus support flows are needed.

EXPLORE connect6

Identify Actionable Retention Risk and Measure What Changes the Outcome

Review one policy-retention use case and determine whether the outcome, data, actionability, intervention, holdout design, economics, and governance support a Model6 pilot.