Sources and Verification (v4)
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
- Primary sources only — official provider pages, not blogs or aggregators.
- Record the "verified on" date — the Microsoft AI certification portfolio is rotating heavily in 2026.
- Separate fact from recommendation — the existence and scope of an exam is a verifiable fact; sequence and timing are judgment.
- Re-verify before every exam — with exams in beta/GA rolling status during 2026, status changes month to month.
Last verified: 2026-07-17
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)
| Certification | Code | Status (2026) | Primary source |
|---|---|---|---|
| Azure AI Apps and Agents Developer Associate | AI-103 | Replaces AI-102 (which retires Jun 30, 2026); validates Python + Microsoft Foundry | Credential page · Study guide |
| Azure Solutions Architect Expert | AZ-305 | Current (Expert level) | Credential page |
| Cloud and AI Security Engineer Associate | SC-500 | New/updated 2026 — verify beta/GA before scheduling | Credential page |
| GitHub Copilot | GH-300 | Current · $99 USD · 700/1000 · valid 24 months · available in Spanish | Credential page · Study guide |
| Azure AI Cloud Developer Associate | AI-200 | New 2026 (beta → GA transition) — verify availability | Official credential announcement |
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)
| Framework | Issuer | Official identifier | Primary source |
|---|---|---|---|
| Azure Well-Architected Framework | Microsoft | WAF | learn.microsoft.com |
| Cloud Adoption Framework | Microsoft | CAF | learn.microsoft.com |
| AI Risk Management Framework | NIST (U.S.) | AI RMF 1.0 (NIST AI 100-1) | nist.gov |
| AI management system | ISO/IEC | ISO/IEC 42001:2023 | iso.org |
| EU AI Act | European Union | Regulation (EU) 2024/1689 | eur-lex.europa.eu |
| Zero Trust Architecture | NIST | SP 800-207 | csrc.nist.gov |
| Top 10 for LLMs | OWASP | OWASP Top 10 for LLM Apps | owasp.org |
| Threats to AI systems | MITRE | MITRE ATLAS | atlas.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.
| Resource | Provider | Cost | Primary source |
|---|---|---|---|
| Azure AI Foundry Documentation | Microsoft | Free | learn.microsoft.com |
| Microsoft Learn (Azure AI / Foundry paths) | Microsoft | Free (badges/Applied Skills) | learn.microsoft.com |
| Official Practice Assessments | Microsoft Learn | Free | learn.microsoft.com |
| DeepLearning.AI — short GenAI/RAG/agent courses | DeepLearning.AI | Free | deeplearning.ai/short-courses |
| Azure Samples (RAG, agents) | Microsoft (GitHub) | Free | github.com/Azure-Samples |
| GitHub Skills | GitHub | Free | skills.github.com |
| Anthropic / OpenAI cookbooks | Anthropic / OpenAI | Free | anthropic.com/learn · cookbook.openai.com |
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
| Before | Now | Note |
|---|---|---|
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 paths | The 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:
| Source | What it contributes | Type | Primary source |
|---|---|---|---|
| WEF Future of Jobs Report 2025 | In-demand skills and AI transformation data (86%, 39%, ~63%) | Research report | weforum.org |
| 70-20-10 model | Experience/social/formal balance in professional development | Framework (CCL) | ccl.org |
| Dunlosky et al. (2013) | Effectiveness of active recall and spaced/retrieval practice | Peer-reviewed study | journals.sagepub.com |
| Google Career Certificates / AWS re/Start / IBM SkillsBuild | Design with projects + capstone + employer connection | Reference programs | grow.google/certificates · aws.amazon.com · skillsbuild.org |
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.