New from the Lab·The Compass — an open moral reasoning standard for AI, tested across frontier modelsExplore →
Production AI Institute · PSF v1.1 open standard
AI Right-To-KnowAI Data Use IndexCheck My AI ToolsPolicy Change WatchAgent ReadinessPublic BenchmarkContactGlobal standard · Worldwide
About How the PSF is built

Built by practitioners, for practitioners

The Production AI Institute is grounded in deployed production systems worldwide. The PSF reflects what breaks in live environments — not what reads well in a generic AI policy.

PAI does not operate a formal external advisory board. Framework development is driven by practitioners with direct production experience — exam feedback, portfolio submissions, ecosystem assessments, and incident analysis all feed the next revision. This page describes that process in practice.

Who we are

The PSF draws on production experience across engineering, operations, governance, and delivery — systems that have run, failed, been fixed, and run again. The framework is the distillation of that experience into something others can use.

Independence is non-negotiable. PAI does not accept sponsorship, affiliate arrangements, or commercial relationships with vendors whose tools we assess. Assessments reflect what we find against published PSF criteria — nothing else.

The institute's mandate is straightforward: publish the definitive open reference for production AI safety and uphold examination and portfolio rigour without diluting published criteria.

How the PSF is developed

The PSF is developed from production experience: real deployments that went wrong, portfolio submissions from CPAP candidates that revealed gaps, and ecosystem assessments that mapped what tools actually cover versus what they leave open.

When a domain is updated, the rationale is published alongside the change. When practitioners identify gaps — through exam feedback, portfolio submissions, or direct contact — those gaps inform the next revision cycle. The framework evolves with the field, not ahead of it.

Examination content is reviewed internally by practitioners with direct production experience in each domain. Scenario questions are drawn from real deployment situations — sanitised and generalised, but grounded in things that have actually happened.

How to contribute

Submit framework feedback

Found a gap, an outdated requirement, or a domain that needs expansion? We review all structured feedback from certified practitioners. Contact us with your credential ID and the specific domain.

Propose examination scenarios

The best exam questions come from real production situations. If you have a scenario that would challenge practitioners — something you've encountered or fixed — we want to hear it. Scenarios are reviewed and rewritten before inclusion to ensure they're educational, not identifiable.

Share production case studies

Real deployments inform better standards. If you have a production case study you're willing to share — what worked, what failed, what the framework missed — we use these to improve the domains and the ecosystem assessments.

Join the assessment panel

As the CPAP and CPAA programmes scale, we are building a panel of practitioners to review portfolio submissions. If you hold CPAP and have breadth of production experience, reach out about joining the review panel.

Get in touch

Questions about the framework, contributing to exam development, or joining the assessment panel?

Contact the PAI team