Skip to content

Details

Most fraud detection talks are code dumps with a toy scenario. Not nice.

Meet Slick Eddie. A fraudster. Five people. Gift cards. And three completely uncoupled transactional systems with zero foreign keys between them. That's the gnarly part nobody talks about.

I wanted to catch him using only ChatGPT-style prompts. No code. Sounded awesome. Then I hit a wall. Hard.

The fraud pattern was just too complex. That's when I partnered Free-Tier Oracle 26ai with an AI Coding Assistant and built a real multi-phase detection pipeline in PL/SQL and PGQL. Fraud "fingerprint" embedding. Similarity search to isolate suspicious transaction bursts. Property Graphs that collapsed 3.7 million paths down to 132. Balance validation to nail the exact fraud sequence.

Is that awesome or what?!

But check this out. This session isn't a victory lap. It's the real story. What I tried. What blew up. And how curiosity plus an AI assistant solved in days what would've taken months solo.

Oracle AI handles every bit of it. Vectors, graphs, SQL. All in one engine. Mind blowing.

Ready to catch Slick Eddie?

Presented by Craig Shallahamer, Applied AI Scientist, Viscosity Founder, OraPub, Oracle ACE Director

This event is co-hosted by the New York Oracle Users Group (www.nyoug.org) and Oracle Professional Services firm, Viscosity North America (www.viscosityna.com)

REGISTER HERE: https://viscosityna.com/how-oracle-ai-tech-and-curiosity-cracked-the-nut-nyoug

You may also like