Search everything.
News, full research text and tables, official policy links, the source directory, and glossary definitions. Research matches link to the newest matching version, including retained historical versions.
Includes your Deep dives date, sorting and reading-time filters. Saved searches rerun against the current library when opened.
Saved searches
Loading saved searches…
8 matches for “AI governance” in Deep dives.
Deep dives
Alloy: distinguish fraud models, decision workflows and AI assistants
A practical evaluation of Fraud Signal, custom-model hosting and agent-assisted investigations, with separate evidence standards for prediction and workflow automation.
Source
[2] Alloy, Actionable AI product documentation; reviewed September 27, 2026; vendor claims https://www.alloy.com/actionable-ai
BioCatch: behavioral signals, scam detection and the limits of a risk score
How behavioral intelligence can complement identity and transaction controls, why unusual behavior is not proof of fraud, and how to evaluate newer sequence-model research.
A behavioral language model is not a chatbot
…investigative visualization and a newer sequence-model concept may sit under the same AI narrative but have different maturity and validation evidence. Procurement should identify which version and function would actually be deployed, what inputs it requires and which outputs are contractually suppor…
Feedzai
Fraud decisioning, digital trust and the evidence needed to evaluate performance claims.
Evidence gates · Table row
Governance · Reason codes, versions, overrides, fairness and monitoring · Challenge, audit and controlled change
Model-risk management after SR 11-7: what SR 26-2 changes
The April 17, 2026 interagency guidance supersedes SR 11-7 and SR 21-8, emphasizes materiality and excludes generative and agentic AI from its formal scope without removing broader governance responsibilities.
Revision summary
…supersedes SR 11-7 and SR 21-8, emphasizes materiality and excludes generative and agentic AI from its formal scope without removing broader governance responsibilities.
Resistant AI: document forensics before an underwriting decision
What document-authenticity models can detect, how their verdicts differ from verified income, and how to handle benign edits, forged statements and customer review paths.
Source
[1] Resistant AI developer documentation, About Resistant Documents; reviewed September 27, 2026 https://developers.resistant.ai/getting-started/about
Roll rates and cures: reading the movement behind delinquency
How transition matrices reveal credit deterioration, why cures need their own definition, and how to avoid mistaking portfolio growth or re-aging for better performance.
Forecasting requires more than matrix multiplication
…different risk from an automated forecast driving material reserves or credit decisions. Governance should follow intended use and consequences, not whether the tool has an AI label.
Sardine: device intelligence and transaction-sequence models for fraud
An evidence-focused review of device signals, issuing-risk models and foundation-model claims, including how to interpret AUC-PR and test cross-institution performance.
Source
[3] Sardine AI Labs, model approach and reported evaluation results; undated page reviewed September 27, 2026 https://www.sardine.ai/ai-labs
Taktile: governing rules, predictive models and AI agents in one decision platform
What Taktile publicly describes, what the evidence does not establish, and how a bank can test decision quality before scaling automation.
What the product actually does
…describes a platform combining a decision engine, data orchestration, case management and an AI Agent Manager. Its public materials distinguish low-code rules and model orchestration from generative assistance: an AI copilot helps write or debug logic, while agents perform configured tasks within workflows.…