PAI credentials are the only vendor-neutral certifications anchored to the Production Safety Framework — a published, independently governed standard for production AI deployment.
Eight credentials across three tracks — examinations (AIDA/AIMA), portfolio pathways (CPAP/CPAA), and specialist examinations (CAIS/CAIG/CLLO/CALLS).
A 20-question scenario-based examination covering all 8 PSF domains. Tests applied knowledge of production AI safety principles. Results are immediate. No prerequisites.
The AIDA equivalent for non-technical leaders. Covers AI governance, risk management, vendor assessment, and regulatory compliance. For product managers, delivery leads, and AI governance roles.
Submit a portfolio documenting a real production AI deployment assessed against PSF criteria. Reviewed by a PAI assessor within 5 business days. Evidences hands-on production competence.
Extended portfolio submission assessed by PAI reviewers. Requires CPAP. The senior PAI credential for principal engineers and practice directors.
Advanced examination covering AI red-teaming, prompt injection defence, output monitoring, multi-agent trust propagation, and safety evaluation frameworks. For practitioners building or operating safety-critical AI systems.
Covers EU AI Act compliance, NIST AI RMF, ISO 42001, algorithmic accountability, and board-level AI governance. For governance leads, legal, risk, and compliance professionals. Exam in development — target Q3 2026.
Practical examination covering LLM fine-tuning, RAG architecture, evals frameworks, inference optimisation, and cost management. For ML engineers and platform teams deploying large language models at scale. Exam in development — target Q4 2026.
The first certification of its kind anywhere. Covers AI discoverability strategy, llms.txt implementation, model context protocol, AI search behaviour, and building brand presence in an AI-mediated information landscape. Exam in development — target Q2 2026.
Independent recognition that an organisation's AI practice meets the PSF standard.
For organisations whose business is building and operating AI systems for clients. Requires ≥3 AIDA holders, at least one CPAP holder, and a review of client deliverables.
For enterprises that deploy AI at scale internally. Requires a documented AI governance framework aligned to PSF and annual self-assessment audit with PAI review.
For employers specifying PAI in job descriptions and procurement requirements, see the Employer Recognition Guide.
All resources are free. AIDA candidates typically use the study guide and practice examination before sitting.