Product
Briefs

Introduction This manual is a practical guide to model risk management for agentic AI, written for model risk teams, compliance officers, and developers. It is intended to help developers and users of agentic systems, including model governance teams, compliance officers, and other financial institution personnel, strengthen their model risk management (MRM) practices by addressing the…

Introduction Historically, there have not been efficient ways to create Less Discriminatory Alternative (LDA) models. Today, because of access to cheap computation and massive amounts of data, it is possible to generate many variations of models, some of which will look more or less fair and accurate than others. As a consequence, lenders need a…

Basic Fairness Metrics Used in Lending and Insurance The federal financial regulators, courts, and industry practitioners use a wide range of fairness measures and thresholds to perform bias testing and fairness analytics. This e-book is intended to provide a non-technical, conceptual explanation of the key fairness measures used in the lending and insurance ecosystem. Specifically,…

Analyzes a business, recommends credit, drafts the memo.

What it takes to trust an agent that handles employee benefits, complaints, accommodations, and sensitive data

What it takes to trust a voice agent that authenticates callers, discloses account information, and moves money

What it takes to trust an agent that decides whether a sanctions alert is a true hit

What it takes to trust an agent that identifies and escalates PEP risk

What it takes to trust an agent that classifies fraud, starts statutory clocks, and handles money movement

What it takes to trust an agent that grants, denies, or executes fee waivers

What it takes to trust an agent that investigates credit-reporting disputes

What it takes to trust an agent that decides whether a consumer issue is a complaint

What it takes to trust an agent that places collection calls

What it takes to trust an agent that talks to consumers about debt

What it takes to trust an agent that reads negative news and decides what matters