Fintech & paymentsStory & newspaper

Limit Cut Autopsy

Every limit-cut release gets a pre-mortem, then an autopsy from the customers it hit.

Listen
Reason
Act
Verify
Learn
Closes 3 of 5: Listen, Reason, Act

The job

When a credit model or policy change cuts limits or declines purchases at checkout, predicts how customers will react before release, hears what they actually say afterwards, and decides against rules set in advance whether it was a bug, badly explained, or fine.

The moment

Example, illustrative figures
  1. Each change that cuts customers' credit limits gets a sealed prediction before release and a review of what customers said after it. The review of release 7 shows 3,100 complaints saying "I never missed a payment".

  2. They cluster on customers who had paid off early. The prediction was 0.8 complaints per 1,000 limit cuts; the actual rate is 6.2, so the forecast is graded F.

  3. The risk team confirms that early repayment was being read as a default, and fixes it that afternoon.

  4. The review of the next release finds it fine, and its forecast is graded A.

What it does

  • Register the pre-mortem
  • Send the evidence note to risk
  • Draft bot, agent and pre-notice copy
  • Stamp the verdict and sign off
  • Drill any cause of death

What you see

Forensic autopsy report (cause-of-death sections, verbatims as exhibits, verdict stamp) preceded by a sealed pre-mortem, plus a committee release register

  1. Screen 1: Pre-mortem
  2. Screen 2: Autopsy
  3. Screen 3: Prescription
  4. Screen 4: Release Register

What it moves

Complaints per 1,000 limit cuts in each release, the time from release to a verdict, and how far off the model team's forecast was.

Built for

  • Risk & fraud
  • Product & digital
  • Compliance, complaints & legal
  • Frontline staff
OnClarity

Bring the one question your current platform can't answer.

Our team will map it to the apps that answer it, then set them up on your own feedback.

Talk to our team

Examples use illustrative figures. Apps are set up with our team on your own feedback data, and setup varies by app.

OnClarity

Bring the one question your current platform can't answer.

Our team will map it to the apps that answer it, then set them up on your own feedback.

Talk to our team

Examples use illustrative figures. Apps are set up with our team on your own feedback data, and setup varies by app.

OnClarity

Bring the one question your current platform can't answer.

Our team will map it to the apps that answer it, then set them up on your own feedback.

Talk to our team

Examples use illustrative figures. Apps are set up with our team on your own feedback data, and setup varies by app.