Use this cheat sheet for Microsoft Certified: AI Business Professional (AB-730) after you understand basic Microsoft 365 workflows. The exam is not about building AI apps. It tests whether you can use generative AI productivity tools, including Microsoft 365 Copilot, Researcher, and Analyst, to improve business work while protecting data, checking output, and keeping a human accountable.
AB-730 answer sequence
flowchart TD
S["Scenario"] --> G["Business goal and audience"]
G --> C["Prompt context, source material, and constraints"]
C --> O["Output type: email, report, presentation, analysis, image, or summary"]
O --> R["Review facts, tone, sensitive data, and policy"]
R --> A["Act with human accountability"]
If an answer asks Copilot to produce important business output without context, source checks, review, or privacy awareness, it is usually weaker.
Assessed area compression
| Assessed area |
What to remember |
Reject when |
| Generative AI fundamentals |
Know prompts, context, hallucination, grounding, responsible use, and model limits. |
AI output is treated as guaranteed truth. |
| Prompts and conversations |
Improve prompts with goal, role, context, constraints, examples, and format. |
The user restarts vaguely instead of refining the conversation. |
| Business content with AI |
Draft, analyze, summarize, create, and refine business work with review. |
AI-generated content is published without checking facts, tone, and policy. |
Business prompt chooser
| Need |
Strong prompt pattern |
Weak prompt |
| draft an email |
audience, goal, tone, source facts, action requested, length |
“write an email” |
| summarize a meeting |
decision points, owners, risks, deadlines, unresolved questions |
“summarize everything” |
| analyze business data |
source, metric definitions, comparison period, decision criteria |
“find insights” |
| create a presentation |
audience, objective, key sections, speaker notes, constraints |
“make slides” |
| generate an image |
purpose, style, content constraints, brand/safety rules |
“make something impressive” |
| refine a document |
target reader, issues to fix, tone, format, unchanged sections |
“make it better” |
Scenario eliminations
| Stem clue |
Eliminate first |
Keep in play |
| confidential business data |
paste into any tool without policy check |
approved tool, least data, access control, review |
| factual report or executive summary |
trust first draft |
source check, citations or source trace, human review |
| biased or sensitive decision |
fully automated final decision |
human accountability, policy, fairness, transparency |
| prompt result is poor |
abandon AI or start vaguely |
refine context, constraints, examples, and output format |
| message has wrong tone |
regenerate randomly |
specify audience, tone, purpose, and constraints |
| spreadsheet analysis |
ask for “insights” only |
define metrics, timeframe, assumptions, and decision needed |
| generated image or content |
publish immediately |
check rights, brand fit, policy, inclusiveness, and accuracy |
Microsoft 365 workflow cues
| App or tool context |
Exam instinct |
| Outlook |
Draft, rewrite, summarize, and adjust tone, but verify recipients and sensitive content. |
| Word |
Draft, revise, summarize, and restructure documents with source-aware review. |
| Teams |
Summarize meetings, action items, decisions, and unresolved questions. |
| PowerPoint |
Build or refine presentation structure for a defined audience and outcome. |
| Excel |
Analyze, summarize, explain, and compare data while checking definitions and assumptions. |
| Researcher |
Gather and synthesize source-aware information for reports. |
| Analyst |
Explore data and reasoning support, but validate conclusions before decisions. |
Responsible business AI rules
| Risk |
Strong answer |
| hallucination |
check source material, ask for evidence, and verify important claims |
| privacy |
minimize sensitive data and use approved tools and access boundaries |
| bias |
review output for unfair assumptions, stereotypes, or missing perspectives |
| accountability |
keep a human owner for final decisions and high-impact actions |
| transparency |
disclose AI involvement where policy or audience expectations require it |
| security |
avoid exposing confidential data, secrets, credentials, or restricted documents |
Final 15-minute review
| If the stem says… |
Start with |
| “best prompt” |
goal, audience, context, constraints, output format, review criteria |
| “business report” |
sources, assumptions, structure, evidence, human review |
| “sensitive data” |
approved tool, minimization, permissions, policy |
| “poor response” |
refine prompt and context instead of trusting or discarding blindly |
| “decision-making” |
AI supports analysis; human remains accountable |
| “Microsoft 365 app” |
choose the Copilot workflow that fits the business artifact |