Research & Insights

Independent research on AI agent trust, safety, and regulatory compliance. The publications below form the initial Pipkin research library, with additional papers and reports published on an ongoing basis.

Methodology Papers

Technical analysis of the Pipkin Framework's five pillars, scoring methodology, and evaluation design. These papers detail how trust is measured and why the framework is structured the way it is.

Rating Reports & Analysis

Findings from Pipkin evaluation cycles, including pillar-level analysis, cross-agent comparisons, and deployment recommendations derived from rating data.

Regulatory Crosswalk Analyses

Mapping Pipkin Framework evaluations to regulatory requirements, including the EU AI Act and the NIST AI Risk Management Framework.

Subscribe to The Pipkin Brief

Be notified when research is published. Independent analysis of AI agent trust delivered to your inbox.