Fintech Risk

Fraud Typology Deep-Dive Prompt

Prepare a fraud typology deep-dive report defining the typology precisely, documenting observed patterns, quantifying detection gaps, and recommending controls.

Anselm Fowel / Public collection2 Oct 2026

Your prompt

The prompt

You are a VP of Fraud Risk who has built fraud typology taxonomies at two fintechs and has personally reverse-engineered a fraud ring's exact application pattern from eleven seemingly unrelated declined applications.

My company has observed a rise in [FRAUD TYPE, e.g. synthetic identity fraud] within [PRODUCT TYPE], representing [LOSS AMOUNT/RATE] over the past [TIME PERIOD], and the fraud team needs a deep-dive to understand the typology before proposing new controls.

First, define the specific typology in precise terms, distinguishing it from adjacent typologies, such as separating synthetic identity fraud from first-party fraud and traditional identity theft, since each requires a different control response.

Next, document the specific pattern of behavior associated with this typology as observed in our data, including application characteristics, velocity patterns, and any device or network signals that recur across confirmed cases.

Then, quantify the current detection rate for this typology, specifying what share of cases are caught before loss versus discovered only after charge-off, and identify the specific point in the customer lifecycle where detection currently fails.

Also, draft the recommended control enhancements targeting this typology specifically, such as a new velocity rule or an identity verification step, with the expected detection lift and any customer friction trade-off.

Finally, synthesize everything into a single Fraud Typology Deep-Dive report with a pattern summary, a detection gap analysis, and a control recommendation table, framed as a first draft for the fraud risk committee to review before new controls are prioritized and built.

1,695 characters

Get new prompts in your inbox

New additions to the prompt library, roughly weekly. No spam, unsubscribe anytime.

Your prompt

Get new prompts in your inbox

Recent articles