AI, ML & Fraud Rules

This isn't a buyer's guide to AI fraud tools, and it isn't a pitch. It's the working notes from someone who actually runs these experiments - testing whether an LLM can score fraud, rebuilding rules that stopped working, and figuring out why a model that looks great in backtesting falls apart in production.

If you're a practitioner trying to get more out of your own AI/ML setup - better rules, fewer false positives, a clearer read on when a model needs a refresh - this is the free version of that thinking. No framework to buy, no demo to sit through.

Learn more in the following articles:

Newsletter

#85 - Stop letting fraudsters write your rules

#84 - The AI hangover no one's preparing for

#80 - How adding friction increased card spend by 11%

#79 - The ML score cheat code almost no one uses in rules

#78 - Are you actually using your ML score?

#77 - AI will fix fraud - just not how you think

#76 - Cut costs with AI - without losing your team or your CFO

#73 - Your fraud system is dying

#72 - The self-learning fraud model is a lie

#71 - The fraud you blocked will break your next model

#70 - You might be testing your solutions blindly

#67 - Your $50 decline just cost you $5,050

#66 - Are you using these 3 rule types?

#64 - 3 Agentic Commerce questions nobody's asking

#63 - Why sophisticated fraud defenses fail more often

#59 - The AI mistake fintech leaders keep making

#57 - Stop trying to predict fraud. Do this instead.

#51 - I gave Claude a hint. Results shocked me.

#50 - The overlooked truth about false positives

#48 - Why LLMs will never replace ML models

#43 - False positives cost you less than you think

#42 - 5 ways to build fraud models without perfect data

#38 - Don't refresh your fraud model until you read this

#35 - When your rules look great but performance tanks

#29 - Beyond AI vs Rules: The 3-Layer Cake of Fraud Defense

#25 - How to Build Fraud Rules that Outperform AI Models

#22 - I asked Claude to score fraud AGAIN. Then this happened...

#13 - The 4 hidden drivers behind false positives

#05 - Claude vs. Fraud: A 3-Hour Game-Changing Test

#04 - The 5 ways to spot False Positives

Working through a similar AI/ML or rules problem and need help?

No pitch, just an outside read on what you're testing.

Book a Free Discovery Call