AI checklists
Model selection, data, evaluation, RAG, agent systems, guardrails, and monitoring.
22 actionable reference checklists. Pick what applies to your project; these are not universal requirements or certifications.
CHECKLISTRequirements and Planning ChecklistVerify that the project solves a validated problem, has measurable goals, clear scope, testable requirements, realistic delivExplore checklist →CHECKLISTLegal Issues ChecklistIdentify legal, licensing, contractual, privacy, and regulatory questions early. This is an issue-spotting tool, not legal adExplore checklist →CHECKLISTAI Evaluation and Governance ChecklistReview whether a proposed AI use is appropriate, its risks are understood, the system has credible evaluation evidence, humanExplore checklist →CHECKLISTSecurity ChecklistAssess confidentiality, integrity, availability, abuse resistance, and operational security proportionate to project risk.Explore checklist →CHECKLISTAI Model Selection and Benchmarking ChecklistChoose AI models against the actual application workload, operating constraints, safety requirements, and measurable outcomesExplore checklist →CHECKLISTAI Data Curation and Labeling ChecklistVerify that datasets used for AI training, tuning, evaluation, or retrieval are lawfully and responsibly sourced, technicallyExplore checklist →CHECKLISTAI Model and Dataset Cards ChecklistVerify that model and dataset documentation accurately communicates intended use, capabilities, limitations, provenance, evalExplore checklist →CHECKLISTAI Fine-tuning and Training ChecklistVerify that training or fine-tuning a model is justified, reproducible, permissioned, numerically stable, evaluated for unintExplore checklist →CHECKLISTPrompt Engineering and Versioning ChecklistReview instructions, templates, prompt revisions, and evaluation gates used in model-integrated software. Treat prompt wordinExplore checklist →CHECKLISTRAG and Knowledge Bases ChecklistVerify that retrieval-augmented generation (RAG) uses authorized, attributable, sufficiently fresh evidence; retrieves relevaExplore checklist →CHECKLISTAI Agents and Tool Orchestration ChecklistVerify that AI agents achieve bounded tasks through permitted tools, safe handoffs, controlled state, and meaningful human suExplore checklist →CHECKLISTAI Inference and Serving ChecklistVerify that production AI inference is deployable, secure, performant, cost-bounded, and recoverable under realistic workloadExplore checklist →CHECKLISTMultimodal AI Quality ChecklistVerify that AI features using images, audio, speech, video, documents, or combinations of modalities produce fit-for-purpose,Explore checklist →CHECKLISTAI Human Review and Escalation ChecklistEnsure human involvement in AI-assisted decisions is meaningful, authorized, timely, and auditable. A review button, nominal Explore checklist →CHECKLISTAI Evaluation and Regression ChecklistVerify that AI-enabled applications have representative, measurable, repeatable evaluation of task quality and safety, and thExplore checklist →CHECKLISTAI Red Teaming and Guardrails ChecklistVerify that AI-enabled products have risk-based adversarial coverage and enforceable defenses against harmful outputs, unauthExplore checklist →CHECKLISTMLOps and Model Lifecycle ChecklistVerify that AI/ML experiments, data, models, approvals, deployment configurations, and retirement decisions are traceable andExplore checklist →CHECKLISTAI Monitoring and Drift ChecklistVerify that production AI systems are observable, that distribution changes are separated from demonstrated model-quality regExplore checklist →CHECKLISTCost and FinOps ChecklistVerify that technology spending is attributable, visible, forecastable, and optimized for business value without compromisingExplore checklist →CHECKLISTAnalytics and Product Metrics ChecklistVerify that product metrics answer meaningful decisions and are built on well-defined events, trustworthy data, privacy-awareExplore checklist →CHECKLISTGEO and AEO ChecklistImprove how clearly and reliably a project can be understood, sourced, and answered by search answer systems and generative AExplore checklist →CHECKLISTRelease and Launch Readiness ChecklistVerify that a new product, service, or significant change is ready for controlled customer exposure, has an explicit go/no-goExplore checklist →