5.1. Exam Strategy and Time Management
Microsoft role-based exams test judgment under constraints. A typical item describes an organization, states one or more requirements, and asks for the option that meets all of them — often with a second option that meets most of them. Read every question in three passes:
- Find the requirement words. "Must," "only," "without," "minimize administrative effort," "least privilege," "before users are affected." These are what separate two plausible answers.
- Classify the decision. Is it about identity, authentication and access, threats, data, Copilot or agent configuration, or a workload? That usually tells you the portal — and wrong answers often name a real setting in the wrong portal.
- Eliminate on one fact. Most distractors fail on a single fact: a retired feature (Targeted release, the legacy ID Protection risk policies, the MSOnline module), a licensing gap (Plan 1 vs Plan 2, Copilot license for federated connectors), or a scope error (RCD on OneDrive, Conditional Access "All users" for agent users).
Time management. Microsoft doesn't publish AB-650's length on the exam page; associate exams typically run 100 minutes. Budget roughly one to one-and-a-half minutes per standard question and bank time for case studies, which carry a long scenario you'll reuse across several questions. Use mark for review freely: answer every question on the first pass (there's no penalty for guessing), flag anything you're less than sure of, and come back once you've seen the whole exam — later questions sometimes jog your memory for earlier ones. Case studies and some question sets lock once you leave them, so finish those before moving on.
Question formats to expect:
- Multiple choice and multiple response — for multiple response, the stem tells you how many to pick; each pick usually has to be independently correct.
- Drag-and-drop and ordering — usually sequences (the steps to verify a domain, to publish an agent) or matching (which portal or role owns each task).
- Hotspot — choose values from dropdowns inside a statement, configuration or PowerShell command; each dropdown is a separate decision, but the item is scored as a whole.
- Case studies — a multi-tab scenario (requirements, existing environment, constraints) with several questions; reread the requirements tab for each question.
- Repeated-answer sets — the same answer list for several scenarios; an answer can be right for more than one, or none.
Preview features. Microsoft says most questions cover generally available features, but preview features can appear if they're commonly used. When an option depends on a feature you know is in preview — for example blocking prompts that contain sensitive information types in the Copilot DLP location — treat it as plausible rather than impossible.
Recency traps. This exam replaced MS-102, so much published study material predates the changes below. If an option relies on the old behavior, it's almost certainly a distractor.
| Old (don't pick) | Current |
|---|---|
| Manage MFA and SSPR methods on the legacy pages | Authentication methods policy |
| User and sign-in risk policies on the ID Protection pages | Risk-based Conditional Access policies |
| MSOnline and AzureAD PowerShell | Microsoft Graph PowerShell |
| Targeted release | Frontier, Standard and Deferred release |
| Restricted SharePoint Search | Restricted Content Discovery |
| DSPM for AI | DSPM (Solutions > DSPM) |
| Copilot Control System, Microsoft 365 Copilot | Copilot controls, Microsoft Copilot |
| Microsoft Graph connectors | Copilot connectors |