AB-100 Exam Update November 2026: New Domain Weights and How to Re-Plan Your Study
From 16 November 2026 the English AB-100 exam weights Plan at 35-40%, Design at 30-35% and Deploy at 20-25%. Here is what changed, what stayed the same, which outline applies to you and how to move your study hours.
Examinotion Team

Last updated: November 2026. Written and fact-checked by the Examinotion team against the official Microsoft Learn study guide and exam page for AB-100.
TL;DR From 16 November 2026, the English AB-100 exam weights Plan AI-powered business solutions at 35-40%, Design at 30-35% and Deploy at 20-25%. Deploy was previously the largest domain at 40-45%. Microsoft's change log lists no added or removed skill groups, so the update changes where your marks come from rather than what you study.
If you are preparing for AB-100: Agentic AI Business Solutions Architect, the exam you sit on or after 16 November 2026 will not be balanced like the one most study guides describe. Microsoft has published a new skills outline that moves the weight of the exam away from deployment and towards planning and design [1]. This article explains exactly what changed, what stayed the same, which version applies to you, and how to move your remaining study hours so they match the new outline.
What changes on 16 November 2026
Microsoft updates the English version of AB-100 on 16 November 2026, and the change is a re-weighting of the three exam domains. The AB-100 exam page carries the notice "Updates to the AB-100 exam are coming on November 16, 2026" and confirms that "the English language version of this exam will be updated on November 16, 2026" [2]. The study guide now lists the skills measured "as of November 16, 2026" [1].
The headline changes are:
- Plan AI-powered business solutions rises to 35-40% and becomes the largest domain [1]
- Design AI-powered business solutions rises to 30-35% [1]
- Deploy AI-powered business solutions falls to 20-25%, the smallest domain [1]
- The audience profile is marked "No change" in Microsoft's change log [1]
- Three skill groups receive "Minor" updates: designing the overall AI strategy, designing AI and agents, and designing the ALM (application lifecycle management) process [1]
Nothing in the change log is marked "Major". Microsoft's own summary describes the three domains only as "% of exam increased" or "% of exam decreased", and every skill group it lists is the same before and after the update [1]. In other words, the AB-100 syllabus is substantially the same; the scoring emphasis is not.
AB-100 domain weights before and after
The AB-100 update takes roughly 20 percentage points of weight away from Deploy, while Plan gains about 10 points and Design about 5. The table below compares the outline in force until 15 November 2026 with the outline that applies from 16 November 2026.
| Domain | Until 15 November 2026 | From 16 November 2026 | Direction (Microsoft change log) |
|---|---|---|---|
| Plan AI-powered business solutions | 25-30% | 35-40% | % of exam increased |
| Design AI-powered business solutions | 25-30% | 30-35% | % of exam increased |
| Deploy AI-powered business solutions | 40-45% | 20-25% | % of exam decreased |
The new percentages and the direction of each change come from the current Microsoft Learn study guide [1]. Microsoft's change log does not restate the old percentages, so the "until 15 November" column reflects the outline Microsoft published when AB-100 launched, which most candidates and existing study guides still use.
The practical effect is easy to underestimate. Microsoft says most of its certification exams contain between 40 and 60 questions [4]. On a hypothetical 50-question exam, Deploy would have supplied around 20 to 22 questions under the old weights and around 10 to 12 under the new ones, while Plan grows from around 13 to 15 questions to around 17 to 20. These are illustrations based on the published ranges, not a statement of how many questions any individual exam contains.
Plan AI-powered business solutions (35-40%): now the largest domain
The Plan domain tests whether you can turn a business problem into an AI strategy that is justified, governed and costed. It contains three skill groups: analyse requirements for AI-powered business solutions, design the overall AI strategy for business solutions, and evaluate the costs and benefits of an AI-powered business solution [1].
Analyse requirements
The requirements skill group focuses on whether agents and data are fit for the job. Microsoft lists assessing "the use of agents in task automation, data analytics, and decision-making", reviewing "data for grounding, including accuracy, relevance, timeliness, cleanliness, and availability", and organising business solution data so it is available to other AI systems [1]. Expect scenarios where the right answer is to fix the data before building anything.
Design the overall AI strategy
The AI strategy skill group is the largest in the Plan domain, with 13 bullets in the current outline. Examples include "Implement the AI adoption process from the Cloud Adoption Framework for Azure", "Determine when to build custom agents or extend Microsoft 365 Copilot", "Provide guidelines for creating a prompt library", "Develop the use cases for customized small language models for the solution" and "Include the elements of the Microsoft AI Center of Excellence" [1].
The Cloud Adoption Framework (CAF) AI guidance is the natural reference point for this group. Microsoft describes it as an overview that "helps decision makers determine how their organization should adopt AI", and notes that "a central governance function, such as an AI Center of Excellence, can define Responsible AI requirements and common boundaries" [7]. If you have not read the CAF AI adoption pages end to end, this is the single most valuable reading task to add before your exam.
Evaluate costs and benefits
The cost and benefit skill group is short, with four bullets, but each one is a classic architect decision. Microsoft asks you to "Select ROI criteria for AI-powered business solutions, including the total cost of ownership", "Create an ROI analysis for the proposed AI solution for a business process", "Analyze whether to build, buy, or extend AI components for business solutions" and "Implement a model router to intelligently route requests to the most suitable model" [1].
The model router bullet is the most technical item in an otherwise strategic domain. Microsoft Foundry documentation describes model router as "a trained language model that selects the best large language model (LLM) to respond to a prompt in real time", with routing modes that balance quality against cost [9]. For AB-100, understand why an architect would choose it, typically to control cost without sacrificing answer quality, rather than memorising deployment steps.
Design AI-powered business solutions (30-35%)
The Design domain tests your ability to choose and shape the right agents, extensions and prebuilt AI features for a business scenario. It contains three skill groups: design AI and agents for business solutions, design extensibility of AI solutions, and orchestrate configuration for prebuilt agents and apps [1].
Design AI and agents
The AI and agents skill group is the largest single group in AB-100, with 16 bullets. It covers designing "task agents", "autonomous agents" and "prompt and response agents", designing topics for Copilot Studio "including fallback", designing agents for Dynamics 365 Contact Center channels, applying the Microsoft Power Platform Well-Architected Framework to intelligent application workloads, and deciding "when to use standard natural language processing, conversational language understanding, or generative AI orchestration in Copilot Studio" [1].
Copilot Studio sits at the centre of this group. Microsoft describes it as "a graphical, low-code studio for building and managing AI-powered agents and workflows" [8]. If you want a business-level refresher on how agents differ from Copilot chat, our guide to Microsoft Copilot agents explained is a useful primer before you go deeper.
Design extensibility
The extensibility skill group covers how agents reach beyond their default capabilities. Bullets include designing AI solutions with custom models in Microsoft Foundry, designing agents in Microsoft 365 Copilot, "Design agent extensibility with Model Context Protocol in Copilot Studio", "Design agents to automate tasks in apps and websites by using Computer Use in Copilot Studio" and designing agent behaviours "including reasoning and voice mode" [1].
Model Context Protocol (MCP) and Computer Use appear in the current outline, but Microsoft's change log works at skill-group level and does not identify which individual bullets are new [1]. Treat them as examinable either way. If you have never built a Foundry agent, our hands-on guide to building your first agent in Microsoft Foundry gives you practical context for the custom-model bullets.
Orchestrate prebuilt agents and apps
The prebuilt agents skill group tests whether you know when not to build. It asks you to orchestrate AI features in Dynamics 365 apps for finance and supply chain and for customer experience and service, propose Microsoft 365 agents for business scenarios, orchestrate the configuration of Microsoft 365 Copilot for Sales and Microsoft 365 Copilot for Service, and propose Microsoft Power Platform AI features "including AI hub" [1]. These bullets pair naturally with the build, buy or extend decision in the Plan domain.
Deploy AI-powered business solutions (20-25%): smaller, not optional
The Deploy domain halves in weight but keeps all four of its skill groups. It still covers analysing, monitoring and tuning AI-powered business solutions; managing testing; designing the ALM process; and designing responsible AI, security, governance, risk management and compliance [1].
Key bullets that remain examinable include "Monitor agent performance and metrics", "Interpret telemetry data for performance and model tuning", "Recommend the process and metrics to test agents", "Design the ALM process for Copilot Studio agents, connectors, and actions", "Design the ALM process for Microsoft Foundry Agents Service", "Analyze solution and AI vulnerabilities and mitigations, including prompt manipulation" and "Validate data residency and movement compliance" [1].
Do not drop Deploy from your plan. At 20-25%, the domain can still decide a borderline result, and its security and governance bullets overlap heavily with the Plan domain's Centre of Excellence and responsible AI content. The efficient approach is to keep Deploy revision focused on decisions (which test metrics, which ALM path, which governance control) rather than configuration detail.
What has not changed
The AB-100 update leaves the exam's structure, scoring and certification rules intact. The following points are unaffected:
- Passing score: 700 [2]
- Skill groups: Microsoft's change log shows the same skill group names before and after the change and lists no additions or removals, with three marked "Minor" and none marked "Major" [1]
- Certification: AB-100 leads to Microsoft Certified: Agentic AI Business Solutions Architect Expert [3]
- Prerequisite: you must also hold at least one qualifying associate-level certification. Microsoft states that "in addition to passing the AB-100 required exam, you must also earn at least one of the prerequisite Associate-level certifications" [3]. Qualifying certifications include AI-103 (Azure AI Apps and Agents Developer), AB-620 (AI Agent Builder) and AB-410 (Intelligent Applications Builder), alongside AI-300 (Machine Learning Operations Engineer) and several Dynamics 365 and Power Platform associates [2]. AB-730, AB-731 and AB-900 do not qualify [2][3]
- Retake rules: if you do not pass, you wait 24 hours before a second attempt, then 14 days between subsequent attempts, with a maximum of five attempts in the 12 months from your first attempt [5]
- Renewal: expert certifications expire annually and renew through a free online assessment on Microsoft Learn [1]
- Preview features: the study guide still notes that most questions cover generally available features, but the exam "may contain questions on Preview features if those features are commonly used" [1]
If you are still choosing your qualifying associate exam, the AB-620 exam guide covers the Copilot Studio route, which overlaps strongly with the AB-100 Design domain.
Booked before or after 16 November? Which outline applies
If you sit the English AB-100 exam on or after 16 November 2026, prepare for the new weights. Microsoft states that the English version "will be updated on November 16, 2026" [2]. Microsoft does not publish separate guidance for candidates who booked before that date, so the safest reading is that the version you receive depends on the day you sit the exam, not the day you booked it.
Localised versions follow later. Microsoft says localised exams "are updated approximately eight weeks after the English version is updated", and the exam page adds that localised updates may not always land on that schedule [1][2]. Eight weeks after 16 November 2026 falls in mid-January 2027, which is our arithmetic rather than a Microsoft date. AB-100 is offered in English and 12 other languages, including Japanese, German, French, Spanish and Portuguese (Brazil) [2]. If the exam is not available in your preferred language, Microsoft allows you to request an additional 30 minutes [1].
Use the timing to your advantage. If your exam is booked for the days just before 16 November and you have revised to the old outline, sitting as booked is reasonable. If your date is later, or you are booking now, revise to the new weights from today. Microsoft also says it reviews its cloud-based exams quarterly [6], so check the study guide's change log again in the week before your exam.
How to re-plan your AB-100 study hours
The simplest way to adapt is to reallocate your remaining study time in proportion to the new domain weights. The table below shows an illustrative 60-hour plan. Examinotion calculated it from the midpoint of each published range, scaled so each column totals 60 hours.
| Domain | Hours under the old weights | Hours under the new weights | Change |
|---|---|---|---|
| Plan AI-powered business solutions | 17 | 24 | +7 |
| Design AI-powered business solutions | 17 | 21 | +4 |
| Deploy AI-powered business solutions | 26 | 15 | -11 |
| Total | 60 | 60 | 0 |
Practical steps to make the switch:
- Audit your notes against the current study guide. Open the AB-100 study guide on Microsoft Learn and tick off each bullet in the Plan domain first, because it now carries the most marks [1].
- Move the freed Deploy hours to Plan. Spend them on the Cloud Adoption Framework AI adoption guidance [7], ROI and total cost of ownership reasoning, and the build, buy or extend decision [1].
- Strengthen Design with scenario practice. Focus on choosing between task, autonomous and prompt-and-response agents, and on when to use MCP or Computer Use in Copilot Studio [1].
- Keep Deploy, but compress it. Revise monitoring, testing, ALM and governance as decision frameworks rather than step-by-step configuration [1].
- Re-test before you book. Use the practice assessment Microsoft links from the study guide, plus scenario-style practice questions, to check your Plan and Design scores specifically.
For a wider view of preparation strategy, our guide on how to pass the AB-100 exam covers resources and exam technique, and the experiences of early AB-100 candidates show the scenario-heavy style you should expect. If you use any guide written before October 2026, check its domain weights against the table above.
Practising the scenario-heavy Plan questions
Plan questions reward judgement, so practise with scenarios rather than definitions. The Plan bullets ask you to select, create, analyse and determine, which signals case-style questions where several answers are plausible and one fits the business constraints best [1]. Microsoft does not disclose question formats in advance, but you can explore the general exam interface in the exam sandbox [4].
Useful scenario patterns to rehearse include:
- Build, buy or extend: a business needs an AI capability that Microsoft 365 Copilot or a prebuilt Dynamics 365 agent already covers, so building custom is the wrong answer [1]
- ROI framing: choosing ROI criteria that include total cost of ownership, not just licence cost [1]
- Model choice: recognising when a model router or a customised small language model suits the cost and quality constraints [1][9]
- Adoption and governance: identifying when a Centre of Excellence or a responsible AI policy is the missing piece before scaling [7]
Microsoft Learn is available during associate and expert exams, excluding Q&A, practice assessments and your profile, and the timer keeps running while you browse [4]. Know where the CAF AI adoption pages and Copilot Studio documentation sit, but do not plan to look things up for every question.
Our free AB-100 practice questions are a good place to test your reasoning before moving on to full practice tests.
Frequently Asked Questions
When does the AB-100 exam change?
The English version of the AB-100 exam changes on 16 November 2026, when Microsoft's new skills outline takes effect. Localised versions in other languages are updated approximately eight weeks later, which points to around mid-January 2027, although Microsoft notes that localised updates do not always land exactly on that schedule.
What are the new AB-100 domain weights?
From 16 November 2026, AB-100 weights Plan AI-powered business solutions at 35-40%, Design AI-powered business solutions at 30-35% and Deploy AI-powered business solutions at 20-25%. Previously, Plan and Design were each 25-30% and Deploy was the largest domain at 40-45%, so the emphasis has moved from deployment to planning.
Were any new topics added to AB-100 in November 2026?
Microsoft's change log marks three skill groups as "Minor" changes and lists no added or removed skill groups. The change log works at group level, so it does not identify individual new bullets. Topics such as model router, Model Context Protocol and Computer Use are all in the current outline and should be treated as examinable.
Does the change apply if I booked AB-100 before 16 November?
Microsoft has not published separate rules for candidates who booked before the change. Because Microsoft states that the English exam is updated on 16 November 2026, the safest assumption is that the outline depends on the date you sit the exam. If you sit on or after 16 November in English, revise to the new weights.
Has the AB-100 passing score or prerequisite changed?
No. The AB-100 passing score remains 700, and earning the Agentic AI Business Solutions Architect Expert certification still requires at least one qualifying associate certification, such as AI-103, AB-620 or AB-410. AB-730, AB-731 and AB-900 do not count as prerequisites. The certification still renews annually through a free Microsoft Learn assessment.
Should I still study the Deploy domain for AB-100?
Yes. Deploy falls to 20-25% of the AB-100 exam but keeps all four skill groups: monitoring and tuning, testing, ALM, and responsible AI, security and governance. At that weight it can still decide a borderline result, so compress your Deploy revision into decision frameworks rather than dropping the domain entirely.
Which AB-100 domain carries the most marks now?
Plan AI-powered business solutions carries the most marks from 16 November 2026, at 35-40% of the AB-100 exam. It covers analysing requirements, designing the overall AI strategy using frameworks such as the Cloud Adoption Framework, and evaluating costs and benefits, including ROI, total cost of ownership and build, buy or extend decisions.
Conclusion
The 16 November 2026 AB-100 update is a re-weighting, not a rewrite. Plan becomes the largest domain at 35-40%, Design rises to 30-35% and Deploy falls to 20-25%, while Microsoft's change log lists no added or removed skill groups [1] and the passing score and certification rules stay the same [2][3]. The candidates who benefit most will be those who move study hours into strategy, ROI and agent design now rather than discovering the shift on exam day.
When you are ready to test yourself against scenario-style questions, begin your AB-100 preparation with Examinotion's practice tests, or browse Examinotion's exam preparation courses to compare every Microsoft AI certification we cover.
Sources
- Study guide for Exam AB-100: Agentic AI Business Solutions Architect - Microsoft Learn, accessed 2026-11-06
- Exam AB-100: Agentic AI Business Solutions Architect - Microsoft Learn, accessed 2026-11-06
- Microsoft Certified: Agentic AI Business Solutions Architect Expert - Microsoft Learn, accessed 2026-11-06
- Exam duration and exam experience - Microsoft Learn, accessed 2026-11-06
- Exam and Applied Skills assessment lab retake policies - Microsoft Learn, accessed 2026-11-06
- Exam frequently asked questions - Microsoft Learn, accessed 2026-11-06
- AI adoption - Cloud Adoption Framework - Microsoft Learn, accessed 2026-11-06
- Overview - Microsoft Copilot Studio - Microsoft Learn, accessed 2026-11-06
- How to use model router for Microsoft Foundry - Microsoft Learn, accessed 2026-11-06
Related Articles

AB-620 Exam Update December 2026: New Domain Weights, Skills and Workflows Explained
Microsoft updates the English AB-620 exam on 18 December 2026. Planning and configuration rises to 50-55%, skills join the outline, agent flows become workflows and topics leave it. This guide explains every change and how to re-plan your study.

Microsoft 365 Copilot Is Now Microsoft Copilot: What the Rename Means for AB-730 and AB-731 (2026)
Microsoft renamed Microsoft 365 Copilot to Microsoft Copilot, and the AB-730 and AB-731 study guides now tell candidates how to read both names. This 2026 guide explains what changed, what did not and the licensing exception.

AB-900 vs AI-901: Which Microsoft Fundamentals Exam Should You Take First in 2026?
AB-900 is for Microsoft 365 administrators and AI-901 is for aspiring AI engineers who can read Python. This 2026 comparison covers content, coding, difficulty, logistics and what each exam leads to.