Skip to main content

Sources and Verification (v4)

Why this page exists

An architecture/engineering-level plan is only credible if its certifications and resources are real, current, and independently verifiable. This page documents every credential, framework, and key resource in the v4 plan with links to primary sources (Microsoft Learn, GitHub, standards bodies) so anyone — or a technical recruiter — can confirm them directly.

Verification method

  1. Primary sources only — official provider pages, not blogs or aggregators.
  2. Record the "verified on" date — the Microsoft AI certification portfolio is rotating heavily in 2026.
  3. Separate fact from recommendation — the existence and scope of an exam is a verifiable fact; sequence and timing are judgment.
  4. Re-verify before every exam — with exams in beta/GA rolling status during 2026, status changes month to month.

Last verified: 2026-07-17

The Microsoft AI portfolio is changing in 2026

Microsoft announced a renewal of Azure and Data/AI credentials. Several exams in this plan are new or transitioning from beta → GA during 2026. Confirm status (beta/GA), skills measured, and practice assessment availability on the official page before scheduling. Official announcement: Updates to several Azure and Data/AI certifications.


v4 plan certifications (verified)

CertificationCodeStatus (2026)Primary source
Azure AI Apps and Agents Developer AssociateAI-103Replaces AI-102 (which retires Jun 30, 2026); validates Python + Microsoft FoundryCredential page · Study guide
Azure Solutions Architect ExpertAZ-305Current (Expert level)Credential page
Cloud and AI Security Engineer AssociateSC-500New/updated 2026 — verify beta/GA before schedulingCredential page
GitHub CopilotGH-300Current · $99 USD · 700/1000 · valid 24 months · available in SpanishCredential page · Study guide
Azure AI Cloud Developer AssociateAI-200New 2026 (beta → GA transition) — verify availabilityOfficial credential announcement
Context for the AI-102 → AI-103 transition

The official page confirms that AI-102 retires on June 30, 2026 and that AI-103 (Azure AI Apps and Agents Developer Associate) is its successor, focused on building generative AI apps and multi-agent solutions with Microsoft Foundry using Python. There is no automatic migration: the new credential requires taking the new exam. Other relevant 2026 portfolio moves: AI-901 (replaces AI-900), AI-300 (MLOps Engineer). Always confirm at the source.


Architecture, governance, and security frameworks (verified)

FrameworkIssuerOfficial identifierPrimary source
Azure Well-Architected FrameworkMicrosoftWAFlearn.microsoft.com
Cloud Adoption FrameworkMicrosoftCAFlearn.microsoft.com
AI Risk Management FrameworkNIST (U.S.)AI RMF 1.0 (NIST AI 100-1)nist.gov
AI management systemISO/IECISO/IEC 42001:2023iso.org
EU AI ActEuropean UnionRegulation (EU) 2024/1689eur-lex.europa.eu
Zero Trust ArchitectureNISTSP 800-207csrc.nist.gov
Top 10 for LLMsOWASPOWASP Top 10 for LLM Appsowasp.org
Threats to AI systemsMITREMITRE ATLASatlas.mitre.org

Free and low-cost technical resources (trusted sources)

Verified stack for building the plan's technical evidence (RAG, agents, evaluation) at no cost or low cost.

ResourceProviderCostPrimary source
Azure AI Foundry DocumentationMicrosoftFreelearn.microsoft.com
Microsoft Learn (Azure AI / Foundry paths)MicrosoftFree (badges/Applied Skills)learn.microsoft.com
Official Practice AssessmentsMicrosoft LearnFreelearn.microsoft.com
DeepLearning.AI — short GenAI/RAG/agent coursesDeepLearning.AIFreedeeplearning.ai/short-courses
Azure Samples (RAG, agents)Microsoft (GitHub)Freegithub.com/Azure-Samples
GitHub SkillsGitHubFreeskills.github.com
Anthropic / OpenAI cookbooksAnthropic / OpenAIFreeanthropic.com/learn · cookbook.openai.com
For someone between jobs

The free stack (Microsoft Learn + Azure AI Foundry docs + Azure Samples + DeepLearning.AI + official practice assessments) covers almost all evidence-building. The mandatory expense is the certification exams; Microsoft periodically offers free vouchers through Virtual Training Days. Check the Microsoft Learn events page before paying.


Product changes already reflected

BeforeNowNote
Azure AI Studio (/azure/ai-studio/)Azure AI Foundry (/azure/ai-foundry/)Renamed by Microsoft; URLs updated in Phase 5
Power BI "Guided Learning"Modular Power BI pathsThe linear path was retired (Phase 2)
AI-102 (Azure AI Engineer)AI-103 (Azure AI Apps and Agents Developer)AI-102 retires Jun 30, 2026

Learning design frameworks and market data

Methodology and Best Practices v4 is based on these primary sources:

SourceWhat it contributesTypePrimary source
WEF Future of Jobs Report 2025In-demand skills and AI transformation data (86%, 39%, ~63%)Research reportweforum.org
70-20-10 modelExperience/social/formal balance in professional developmentFramework (CCL)ccl.org
Dunlosky et al. (2013)Effectiveness of active recall and spaced/retrieval practicePeer-reviewed studyjournals.sagepub.com
Google Career Certificates / AWS re/Start / IBM SkillsBuildDesign with projects + capstone + employer connectionReference programsgrow.google/certificates · aws.amazon.com · skillsbuild.org
About statistics and proportions

The WEF figures (86% / 39% / ~63%) are employer survey data, not guaranteed projections. The 70-20-10 proportions are a guiding philosophy (CCL, 1988), not a statistically validated formula. Employment outcome percentages from Google/AWS/IBM are reported by each provider.


Note about salary ranges

Any salary range associated with roles in this plan (AI Solution Architect, AI Success Engineer, AI Platform Engineer) is directional, not a guarantee. It varies strongly by country, industry, company size, and seniority. For live figures:

  • Levels.fyi — tech compensation by level and company
  • Glassdoor — ranges by role and location
  • Robert Half / Michael Page reports (verify the year)

Honesty log

  • Verified facts: existence and scope of AI-103, AZ-305, SC-500, GH-300, and AI-200; retirement of AI-102 (Jun 30, 2026); framework identifiers (WAF, CAF, NIST AI RMF 1.0, ISO/IEC 42001:2023, EU AI Act = Regulation (EU) 2024/1689, NIST SP 800-207).
  • Judgment (not fact): the phase sequence, weekly timing, and project selection are recommendations.
  • Moving status: AI-103, AI-200, and SC-500 are transitioning beta → GA during 2026; confirm availability before scheduling.
  • Directional pending verification: salary ranges and any market statistics.