Skip to content

Attrition Prediction Engine reference profile.

Which customers are at risk of churning or canceling, and when?

SaasSubscriptionCustomer SuccessForecast
Share
AI summary

Overview

Gives Customer Success teams advance warning to intervene before a customer churns rather than after.

Who owns this decision

Chief Customer Officer, VP Customer Success

What this Specialist Model actually does

Decision: Which customers are at risk of churning or canceling, and when?

The final taxonomy records 5 supporting or external components, but it doesn't promote an Oumi-owned specialist model for this engine.

Who uses this Specialist Model

  • Chief Customer Officer
  • VP Customer Success
  • VP Revenue Operations
  • Head of Retention

Where the specialist model sits

The final taxonomy records 5 supporting or external components, but it doesn't promote an Oumi-owned specialist model for this engine.

Feature store aggregating usage, billing, support, and survey data

Supporting component

How Oumi connects to your stack →

Retrieval over historical account notes and prior save-play outcomes

Supporting component

See search reasoning models →

Alerting/workflow integration

Supporting component

Alerting/workflow integration with CS platforms

How Oumi connects to your stack →

Tabular/time-series churn-probability model (gradient boosting or similar)

Supporting component

How Oumi connects to your stack →

Specialist LLM

Supporting component

Specialist LLM for risk-driver explanation and CSM-facing summaries

See generation models →

How to evaluate and improve it

Evaluation

  • False positive rate (flagged accounts that renew)
  • Revenue-weighted prediction accuracy
  • Intervention success rate on flagged accounts

Feedback signals

  • Actual renewal or churn outcome
  • CSM-logged intervention and its result
  • Account health score override by a human reviewer
  • Win-back or save-play success/failure
  • Downgrade or contraction event
  • Customer-reported reason for cancellation (exit survey/interview)

Reference profile

Oumi doesn't recommend an Oumi-owned specialist model for this engine's core decision.

Atlas of AI. Recommendations require validation against your data and operating constraints.