Merchant Risk & Chargeback Optimization
This Ospina case study documents how Carlos Rico-Ospina approached a specific risk, infrastructure, revenue, or research problem and what was built to address it.
Cut chargeback rates in half with a hybrid AI + operations system that turned fraud defense into a competitive advantage.
The Problem
In high-risk digital goods processing, maintaining healthy merchant accounts is existential. Chargeback rates were hovering at 1.0–1.2%—the threshold where Visa/Mastercard monitoring programs begin. Standard fraud filters (AVS/CVV) were failing. The company's ability to process payments was at risk.
The Insight
Traditional rules-based fraud prevention was obsolete for digital goods. I identified an early-stage opportunity with a prominent ML fraud detection platform. This shouldn't be a passive vendor relationship—by providing diverse transaction data, we could help train their models for the specific nuances of prop trading while getting enterprise-grade protection that didn't exist for our niche.
What I Built
- Integrated ML-based real-time decisioning using device fingerprinting and behavioral analytics
- Built custom internal dispute database replacing manual email-based chargeback handling
- Created streamlined dashboard for international contractors to access logs and auto-generate evidence packs
- Established mutual value exchange with fraud prevention vendor—data for protection
Outcomes
- Reduced chargeback rate from 1.2% to 0.6% within 60 days
- On $10M monthly volume: $60K/month reduction in disputed transactions
- Moved from 'risk zone' to 'safe zone' with banking partners
- Hundreds of thousands in recovered revenue through automated dispute response
Why It Matters
Treated Fraud Ops as a product problem, not just a finance problem.
Combined contract negotiation with technical architecture—getting ML vendor onboard AND building the ops infrastructure.
Client and vendor names anonymized. Written praise received from payments executives supporting 85,000+ merchants (available upon request).
Related Projects
View All
AI / MLQuantitative Risk Engine
2024Built an ML-driven risk platform that processes 89 million trades and identifies dangerous behavior before blowouts happen.
AI / MLWeather Market Probabilistic Forecasting
2025Built two probabilistic models that identify conditions where forecast skill beats market consensus—turning calibrated edge into directional trading signals.
Have a Similar Challenge?
Let's discuss how I can help you achieve similar results.
Start a Conversation