Show HN: Logic, Inc. – Automate human-in-the-loop fuzzy decisions

How this works

This automation evaluates payment disputes and assigns a risk score (0-100) based on multiple factors. Disputes are then automatically approved, flagged for manual review, or denied based on configurable thresholds.

Risk factors considered:

• Transaction amount and frequency

• Customer history and account age

• Dispute reason and supporting evidence

• Merchant reputation and dispute patterns

• Time elapsed since transaction

Input format

Use the following inputs:

• dispute ID

• customer ID

• transaction amount (must be >0, otherwise return an error)

• transaction date

• dispute reason

• customer account creation date (used to calculate account age)

• customer's previous number of disputes

• merchant dispute rate (in decimal)

• evidence provided

Decision logic

Risk Score Calculation (0-100):

• High transaction amount relative to customer history: +20

• Recent account (< 90 days): +15

• Multiple previous disputes: +25

• High merchant dispute rate (> 5%): -20

• Strong supporting evidence: -15

• Dispute filed within 7 days: -10

Routing thresholds:

• 0-30 (Low risk): Auto-approve refund

• 31-60 (Medium risk): Flag for manual review

• 61-100 (High risk): Auto-deny, provide appeal instructions

Output

Returns a structured decision with reasoning.