Feature · Churn prediction
Most CS teams find out a customer is leaving the day they ask to cancel. Exeechain's churn prediction is built to surface the same risk weeks earlier, with the three biggest drivers per customer explained in plain English, refreshed daily.
What it is
Every customer gets a 0–100 risk score, recalculated daily. The number is not a black box and it is not a guess: it is a weighted scorecard over the signals Exeechain can see, so every point in it traces to a component you can inspect. A score of 78 means nothing by itself. The value is in the three drivers ranked underneath it, each with its own contribution: login down 62% over fourteen days, NPS dropped from 9 to 6, primary champion moved to a different role at the same company. Those are the three things to talk about in tomorrow morning's save call.
Unlike rules-based scoring (ChurnZero, configurable Vitally, custom Planhat formulas), there is nothing to author or maintain: the components and their weights ship with the product and are tuned centrally, not by you. What changes as you connect more sources is how much the score can SEE - a component with no data behind it abstains rather than guessing, so adding support and CRM data in the first fortnight sharpens the picture by removing blind spots, not by retraining anything.
The AI writes the explanation, not the number. A language model turns those ranked contributions into the sentence a CSM reads, and it is never asked to invent a probability - which is why the words and the score cannot disagree.
How it works
Step 01
Connect Stripe (for MRR + customer list) and drop in a JavaScript tracking snippet for product usage. Optionally connect HubSpot, Intercom or Zendesk, and Mixpanel on any paid plan; Salesforce (Command). Each one adds signals the score can read.
Step 02
Components read billing patterns, login frequency, feature adoption, support ticket sentiment, NPS history, champion role changes, payment failures and renewal proximity. Any component with no data behind it abstains rather than guessing, and says so. First scores in 15 minutes. Full accuracy in 24 hours.
Step 03
Every customer gets a fresh 0–100 score every morning. Each score lists the three biggest drivers in plain English, plus the suggested action: save email draft, executive escalation, downgrade conversation, or expansion outreach.
Why it matters for NRR
Net revenue retention is the single most leveraged metric in SaaS. A few points of NRR compound into a dramatically more valuable company over five years.
Most save outreach happens after the customer has already mentally decided to leave. Surfacing risk while the account is still reachable shifts the conversation from damage control to relationship repair. Exeechain does not publish a lead-time figure, because it has not measured one it would defend: the internal replay put it at one to three days on a 1,300-account book, not the weeks the category likes to claim.
Outreach that lands before the decision to leave converts far better than outreach at cancellation. The whole product is designed to move your first touch earlier on that curve.
Revenue projections weighted by current health scores produce defensible quarterly forecasts. The number you take to your board meeting is the number that materializes.
Related features
Customer health score
AI score with the three drivers explained
QBR automation
Eight hours → thirty seconds
Workflow automation
Custom journeys with goals, A/B splits, approval gates
Retention playbooks
3 prebuilt, plus your own - AI-drafted per customer
Copilot
Plain-English Q&A grounded in your data
Revenue forecast
A 90-day number you can defend
First scores in 15 minutes. Full accuracy in 24 hours. From $299/mo. Daily AI churn scoring with the top three drivers per customer.
Not ready to switch? Run the free 60-second revenue leak scan and see what staying put costs.