The CAIS™ Common Body of Knowledge · Delivered inside Prompt Atlas · Public Verification Registry
The curriculum
the profession wrote.
Every Certified in AI Standards (CAIS™) course begins with the Common Body of Knowledge (CBK) — the published list of skills every certified professional must master. Seven domains. Weighted by a formal Job Task Analysis (a structured survey of real practitioners that determines which skills matter most on the job). Crosswalked to the frameworks your regulator, your auditor, and your board already reference.
Authored by GAISB™. Earned inside Prompt Atlas. Recorded in a registry that does not lie.
Prologue · The Delivery LayerThere is one room where this work gets done.
The curriculum below is authored by GAISB™. It is delivered exclusively insidePrompt Atlas — the official GAISB™ community. No third-party school licenses this program. No partner reseller grants this credential. One author. One room. One path.
Layer 02 · The Path
Prompt Atlas.
The official candidate community of GAISB™. Prompt Atlas is a live, active membership where the entire CAIS curriculum is delivered — self-paced coursework alongside scheduled live cohorts, weekly working sessions, peer review, and panel evaluation of every Build and the Capstone. Membership in Prompt Atlas is part of every CAIS tier.
Delivery
Self-paced AIS-101 through AIS-230 · scheduled live cohorts · recorded masterclasses · working sessions
Submission
All 7 Builds submitted, peer-reviewed, and panel-graded inside Prompt Atlas against published rubrics
Capstone
AIS-400 production Capstone defense scheduled and conducted through the Prompt Atlas cohort calendar
Governance
Operates under GAISB™ standards · delivery only · issuance authority rests with GAISB™
Layer 03 · The Record
Public. Citable. Permanent.
On pass, the CAIS credential is recorded in the GAISB™ Public Verification Registry — the official, publicly auditable record of every CAIS holder. Maintained under public bylaws. Cryptographically signed by the Standards Council. Globally verifiable, counterfeit-resistant, and citable in audit trails, hiring workflows, and regulator filings. The first major AI professional credential designed as a citable instrument from day one.
Issuance
GAISB™ Standards Council · under public bylaws · cryptographically signed
Format
Public registry entry · W3C Verifiable Credentials compatible · non-transferable
Audit Trail
Status changes (CPE, sanctions, revocations) appended with full timestamp history · nothing rewritten
Access
Public lookup · open API · verify.gaisb.org for single-record verification
“A vendor can license a course. A platform can issue a certificate. Neither can author the standard, govern the path, and demand the proof.”
GAISB™ · Curriculum Charter, Article I
Chapter One · The CBKSeven domains. One source of truth.
The Common Body of Knowledge is the single document from which every course, every exam item, and every Build Task rubric is derived. It is published, versioned, and open to public comment through Standards Council review cycles.
D1
Foundations of AI & Strategic Context
The economic, societal, and geopolitical arc of AI. Strategic mental models. How AI reshapes firms, markets, labor, and risk.
D2
Generative AI & Model Architectures
Transformer architectures, training paradigms, RLHF, multimodal systems, model capability evaluation, and system constraints.
D3
Prompt Engineering & System Design for LLMs
Prompt architecture, system messages, context engineering, evaluation harnesses, retrieval patterns, and production-grade LLM systems.
D4
AI Agents & Agentic Workflows
Tool use, planning, memory, multi-agent systems, evaluation of autonomous behavior, containment patterns, and agent safety.
D5
Ethics, Data & Responsible AI
Data provenance, bias, privacy, consent, evaluation of harm, disclosure duties, and the Code of Professional Conduct in practice.
D6
AI Strategy, Transformation & Governance
Board-level governance, AI operating models, risk appetite, policy design, and executing transformation across business units.
D7
Applied AI Innovation & Deployment
Designing, scoping, deploying, and iterating production AI systems. Measuring business and user outcomes. Sunsetting responsibly.
Note on weightings.Domain weights are not opinions. They are derived from the CAIS Job Task Analysis — a structured survey of practitioners, hiring managers, and governance leaders across 12 industries — and validated by the Standards Council on a three-year review cycle. When the practice changes, the blueprint changes. When the blueprint changes, the courses change.
“Teach to the CBK. Test to the CBK. Build against the CBK. Everything else is decoration.”
GAISB™ Standards Council · Curriculum Charter, Article II
Chapter Two · The Core CurriculumSeven Builds. One Capstone. Every one panel-reviewed.
Seven core courses (AIS-101 → AIS-230) each terminate in a Build — a production-grade deliverable reviewed against a published rubric inside Prompt Atlas. AIS-400 is the Capstone defense for Tier III and IV. No lectures without a deliverable. No deliverable without a standard. No credential without a record in the Public Registry.
AIS-101 · CORE · TIER I
The Age of AI & Strategic Mindset
D1
Equips candidates to think clearly about where AI is, where it is heading, and how to make strategic decisions under model uncertainty. Covers the economic and geopolitical arc, foundational mental models, and the operator’s decision lens.
Build Task
Publish a 2,000-word strategic brief assessing an AI disruption vector for a named industry. Reviewed by faculty against the Strategy Rubric.
Framework Alignment
OECD · Human-centred valuesNIST · Govern
AIS-120 · CORE · TIER I
Foundations of Generative AI
D2
Transformer architectures, pretraining and post-training, RLHF, multimodal systems, and the capability/constraint boundary of modern foundation models. Candidates learn to read model cards, interpret evaluations, and distinguish marketing claims from measured behavior.
Build Task
Produce a model evaluation report comparing three foundation models across a defined task suite, with reproducible methodology and documented limitations.
Framework Alignment
NIST · MeasureISO 42001 · A.6.2
AIS-130 · CORE · TIER I → II
Prompt Engineering & System Design for LLMs
D3
The engineering discipline for LLM-powered systems. Prompt architecture, system messages, context windows, retrieval-augmented generation, evaluation harnesses, failure-mode analysis, and the observability patterns required for production.
Build Task
Ship a RAG system with evaluation harness, latency budget, and documented failure modes. Repo + runbook submitted to the Registry-eligible portfolio.
Framework Alignment
NIST · ManageISO 42001 · A.6.2.6EU AI Act · Art. 13
AIS-140 · CORE · TIER II
AI Agents & Agentic Workflows
D4
Building and governing systems that take action. Tool use, planning, memory, multi-agent orchestration, evaluation of autonomous behavior, containment patterns, escalation rules, and the audit trail requirements for agent deployments in regulated contexts.
Build Task
Deploy a multi-step agent with authority boundaries, tool allowlist, audit log, and documented rollback. Reviewed by a Tier III+ faculty panel.
Framework Alignment
NIST · Manage & MeasureISO 42001 · A.9EU AI Act · Art. 9, 14
AIS-160 · CORE · TIER I → II
Ethics, Data & Responsible AI
D5
The operator’s guide to getting this right: data provenance, consent and privacy regimes, bias evaluation, disclosure duties, harm assessment, and the CAIS Code of Professional Conduct applied to real cases. Taught alongside the Ethics Review Board’s published precedents.
Build Task
Conduct and publish an AI Impact Assessment for a deployed or proposed system, following the GAISB™ AIA template and submitting for peer review.
Framework Alignment
NIST · Govern & MapISO 42001 · A.5.2, A.5.3EU AI Act · Art. 10, 27OECD · Transparency
AIS-210 · CORE · TIER II → III
AI Strategy, Transformation & Business Innovation
D6
Board-level AI governance and enterprise transformation. Operating models, AI risk appetite, capability portfolios, vendor diligence, policy design, and the sequencing of change across business units. Built for Operators who will stand in front of a board, a regulator, or a CEO.
Build Task
Deliver an AI Operating Model & Governance Charter for a named organization, including AI policy, risk taxonomy, and 18-month transformation roadmap.
Framework Alignment
ISO 42001 · Clauses 4–10NIST · GovernEU AI Act · Art. 17
AIS-230 · CORE · TIER II
Innovation & Applied AI Foundations
D7
How to move from idea to deployed system, measured outcome, and responsible sunset. Discovery, scoping, safety-by-design, MVP release, measurement, iteration, and decommissioning patterns. Every capstone in the CAIS program is executed through this course’s delivery framework.
Build Task
Ship an AI system into production (or closed beta) with a 30-day outcome log, user feedback record, and documented iteration & sunset plan.
Framework Alignment
ISO 42001 · A.6, A.7NIST · Manage
AIS-400 · CAPSTONE · TIER III → IV
The CAIS Capstone — Architect Review
ALL 7 DOMAINS
The terminal requirement for Tier III (Operator) and Tier IV (Architect). Candidates scope, govern, ship, and defend a production AI system or ecosystem. Minimum 30 days in production. Five-panel Standards Council review. Public Registry listing upon pass. This is where the credential earns its name.
Defense Requirements
Live defense before a five-person panel. Written dossier. Code & policy artifacts. Published outcome report.
Framework Alignment
Full 4-framework crosswalk
Standards instruments backing this curriculum
Every exam question in the CAIS program is written to EXM-2026-01, the published Exam Blueprint & Cut Score Methodology. Every Capstone is scored against CAP-2026-01, the published Capstone Specification & Rubric. Both instruments are in public comment — tier-weighted, inter-rater-reliability-governed, and registry-anchored.
Exam Blueprint →Capstone Specification →Chapter Three · The MappingEvery course. Every domain. Every framework.
The matrix below is the contract. It is what faculty teach to, what Item Writers test to, and what employers receive when they verify a CAIS holder. Nothing outside the matrix is examined.
| Course | D1 | D2 | D3 | D4 | D5 | D6 | D7 | Tier |
|---|
| AIS-101 · Age of AI | ● | ○ | ○ | ○ | ◐ | ◐ | ○ | Tier I |
| AIS-120 · Generative AI | ◐ | ● | ◐ | ○ | ○ | ○ | ○ | Tier I |
| AIS-130 · Prompt Engineering | ○ | ◐ | ● | ◐ | ○ | ○ | ◐ | Tier I → II |
| AIS-140 · Agents & Workflows | ○ | ◐ | ◐ | ● | ◐ | ○ | ◐ | Tier II |
| AIS-160 · Ethics & Responsible AI | ◐ | ○ | ○ | ◐ | ● | ◐ | ○ | Tier I → II |
| AIS-210 · Strategy & Transformation | ◐ | ○ | ○ | ○ | ◐ | ● | ◐ | Tier II → III |
| AIS-230 · Applied Innovation | ○ | ○ | ◐ | ◐ | ◐ | ◐ | ● | Tier II |
| AIS-400 · Capstone | ● | ● | ● | ● | ● | ● | ● | Tier III → IV |
● Primary domain | ◐ Secondary coverage | ○ Not examined in this course
Chapter Four · The CrosswalkWhat the CBK teaches is what your regulator requires.
Every CBK domain is crosswalked to specific clauses of four governance frameworks. When a compliance officer asks “does this person’s training cover ISO 42001 clause 7.2?” — the answer is printed on the certificate.
NIST AI RMF
Govern · Map · Measure · Manage
Every course maps to at least one NIST function. The Capstone defends evidence across all four.
EU AI Act
Articles 9, 10, 13, 14, 17, 27
Risk management, data governance, transparency, human oversight, quality management, and fundamental rights impact.
ISO/IEC 42001
Clauses 4–10 & Annex A
~78% clause-level alignment. Detailed crosswalk published in the CAIS–ISO 42001 Alignment Dossier.
OECD AI Principles
All 5 value-based principles
Human-centred values, transparency, robustness, accountability, inclusive growth — woven through D1, D5, and the Code.
Chapter Five · The Delivery · Inside Prompt AtlasSelf-paced. Cohort-supported. Panel-reviewed. Publicly recorded.
The entire operating cadence of the CAIS program runs inside Prompt Atlas — the official GAISB™ community.
01
Self-paced curriculum, live when you need it
Core courses (AIS-101 → AIS-230) run self-paced inside Prompt Atlas. Live cohorts, working sessions, and faculty office hours run on a published cadence — supplement, not replace.
02
Seven Builds, panel-reviewed
Each of the seven Builds is submitted to Prompt Atlas, peer-reviewed by the cohort, then panel-reviewed by Tier III+ faculty against the published rubric. No rubric, no deliverable, no grade.
03
Published rubrics, no mystery
Every Build has a published rubric with weighted criteria. Candidates see it before they begin. Examiners score against it. The same rubric applies everywhere.
04
Proctored, psychometric exams
Blueprinted to CBK weights. Items drawn from a reviewed bank. Cut scores set by Modified Angoff. Every exam form is psychometrically validated.
05
Public Registry issuance on pass
On successful exam + Builds (and Capstone for Tier II+), the credential is recorded in the GAISB™ Public Verification Registry, cryptographically signed by the Standards Council. Sanctions and revocations are appended publicly with full timestamp history.
06
Global delivery, regional cohorts
Prompt Atlas is virtual-first and worldwide. Regional cohorts in English, Spanish, and Portuguese at launch. Additional languages on a two-year rollout. CPE (40 credits annually) runs inside the same community.
The Standard · The Path · The RecordBe taught what will be tested.
Be tested on what is required.
Be recorded where it cannot be edited.
The CBK is public. The rubrics are public. The frameworks are public. The path runs throughPrompt Atlas. The record lives in theGAISB™ Public Verification Registry. The only thing a candidate buys is the apprenticeship that turns knowledge into evidence — and the evidence that turns into a credential no one can quietly alter.
Authored by GAISB™ · Earned inside Prompt Atlas · Proven by Real Builds