Resources
Guides and practices for reliable AI
A curated collection of resources for teams building, evaluating, and operating enterprise AI systems.
Start Here
Foundations for understanding AI failure.
AI Failure Modes Library
The working catalog of recurring AI failure patterns.
Read →AI Failure Modes
How AI systems fail in production and what to do about it.
Read →Agent Failure Modes
How autonomous agents fail differently from chatbots.
Read →What Is FailureModes.ai?
Our thesis, scope, and how we work with enterprise teams.
Read →FAQ
Common questions about diagnostics, engagements, and how we work.
Read →Practices
Operational disciplines for production AI.
LLM Evals
Design evals that reveal production failure modes and model regressions.
Read →AI Red Teaming
Systematic adversarial testing for enterprise AI systems.
Read →AI Monitoring
Runtime observability and detection for live AI systems.
Read →Model Risk Management
Governance, controls, and risk frameworks for AI at scale.
Read →Library
The AI Failure Modes Library.
Hallucination
Fabricated content presented as fact.
Read →Tool Misuse
Agents calling tools incorrectly or unsafely.
Read →Prompt Injection
Untrusted input hijacking model behavior.
Read →Context Drift
Loss of relevant context across long interactions.
Read →Refusal Drift
Models over- or under-refusing over time.
Read →Cost Runaway
Unbounded token or tool-call spend in production.
Read →Schema Violation
Outputs that break expected structure or contracts.
Read →Cascading Agent Failure
One agent error compounding through a chain.
Read →Evidence
Case studies, benchmarks, and examples.
Community
Contribute and learn from practitioners.
Community Failure Modes
Failure patterns submitted and discussed by practitioners.
Read →Submit a Failure Mode
Share a failure mode you've observed in production.
Read →Featured Submissions
Curated highlights from the community.
Read →Contributor Guidelines
What we look for and how to contribute well.
Read →