---
title: "AB-730 Study Guide — AI Business Professional"
description: "An interactive study guide built on 7 memory techniques to help you pass the Microsoft Certified: AI Business Professional (AB-730) exam — no coding required."
url: "https://lucidlabs.com.au/insights/ab-730"
---

# Master the AB-730

An interactive study guide built on 7 memory techniques to help you pass the Microsoft Certified: AI Business Professional (AB-730) exam — no coding required.

Understand Generative AI Fundamentals ~33%Manage Prompts and Conversations by Using AI ~33%Draft and Analyze Business Content by Using AI ~33%

What it covers

Generative AI basics, using Copilot Chat and Copilot inside Word, Excel, PowerPoint, Outlook and Teams, writing effective prompts, Copilot agents (Researcher and Analyst), notebooks and pages, and responsible AI habits for everyday business use. No coding or app-building skills required. Microsoft hasn’t published official weightings for this exam — treat the three domains as roughly equal, around a third each.

Ideal for

Everyday Microsoft 365 users, business professionals, and team leads who use Copilot to draft emails, build presentations, analyse spreadsheets, and manage documents — not developers, data scientists, or IT admins.

Aspire to this if

You want to work smarter with Microsoft 365 Copilot day to day, help your team adopt AI with confidence, or simply prove your practical AI skills without writing a line of code.

Section 1 / Spatial Memory

## The Map

Tap any component to see what you need to know for the exam.

**🌱 Generative AI Basics**

LLMs · Prompts · New Content

**💬 Copilot Chat**

Standalone Assistant · Web + Work Data

**🛡️ Responsible AI, Every Day**

Check Sources · Protect Data · Stay Accountable

**✍️ Prompt Basics (GCES)**

Goal · Context · Expectations · Source

**🤖 Copilot Agents**

Researcher · Analyst · Multi-Step Work

**🔄 Refining the Conversation**

Follow-Ups · Feedback · Iteration

**📧 Copilot in Word & Outlook**

Draft · Rewrite · Summarise

**📊 Copilot in Excel & PowerPoint**

Analyze · Visualise · Build a Deck

**📓 Notebooks & Pages**

Organise Sources · Save Your Work

Section 2 / Narrative Memory

## The Story

Follow the narrative to build a mental model of how everything connects.

🌱

### Meet Generative AI

Generative AI doesn’t just find information — it creates new content. Underneath Microsoft 365 Copilot sits a large language model, trained to predict the next likely word based on patterns in huge amounts of text. Combine that with your emails, documents, and meetings, and you get an assistant that can draft, summarise, and brainstorm in your own business context.

**Exam Intel**Know the basic idea: generative AI creates new content (text, images, summaries), powered by an LLM predicting likely words. Copilot = generative AI + your work data. AB-730 tests plain understanding, not technical detail.

💬

### Say Hello to Copilot Chat

Copilot Chat is your open-ended AI assistant — not tied to any single document. Open it in a browser, in Teams, or on your phone, and ask it anything from "explain this industry term" to "draft me an announcement about our new product". It can lean on general knowledge or your organisation’s data, depending on what you ask for.

**Exam Intel**Copilot Chat = standalone assistant, used on its own rather than inside a specific app. Contrast with "Copilot in an app" (Word, Excel, etc.), which works on the document you already have open.

🛡️

### Using AI Responsibly, Every Day

Copilot is a drafting assistant, not a decision-maker — you’re still accountable for what goes out under your name. Check citations when Copilot points to a source. Never paste confidential data into a prompt without thinking it through first. And remember: Copilot only ever sees what you’re already allowed to see.

**Exam Intel**Responsible use for business users: verify AI output before acting on it, avoid putting sensitive/personal data in prompts, check citations, and remember the human stays accountable. Copilot respects existing permissions — no data elevation.

✍️

### The Four Ingredients of a Good Prompt

A strong prompt has four ingredients: your Goal (what you actually want), Context (the background Copilot needs), your Expectations (tone, length, format), and a Source (what it should base the answer on). Miss one, and you’ll spend more time editing the result than if you’d included it up front.

**Exam Intel**Memorise GCES: Goal, Context, Expectations, Source. Expect scenario questions asking which ingredient a weak prompt is missing. Specific prompts beat vague ones.

🤖

### When You Need More Than a Quick Answer

Some questions are too big for a single chat reply. That’s where Copilot agents come in. Ask the Researcher agent to dig through your files, meetings, and the web to build a cited report on a topic. Ask the Analyst agent to explore a dataset, test a hypothesis, and explain what’s really driving the numbers — all without you writing a formula.

**Exam Intel**Two named agents to know: Researcher (deep research, citations, across files/email/web) and Analyst (data exploration, trend explanation). Agents plan and investigate over multiple steps, unlike a single prompt-response.

🔄

### Coaching the Conversation

Copilot conversations build on themselves — you don’t need to restate everything each time. If the first draft isn’t quite right, coach it: "shorter", "more formal", "focus on the budget impact". Think of it less like typing a search query and more like briefing a capable colleague.

**Exam Intel**Copilot keeps context within a conversation (multi-turn). Refine with direct feedback rather than starting over. Start a new chat to reset context for a new topic.

📧

### Copilot Where You Already Work

Copilot in Word can turn a blank page into a first draft, tighten a wordy section, or summarise a 40-page report into a page of key points. Copilot in Outlook drafts replies, summarises threads you’ve missed, and can soften — or sharpen — the tone of an email before you hit send.

**Exam Intel**Word Copilot: draft / rewrite / summarise a document. Outlook Copilot: draft replies / summarise threads / coach tone. You always review before sending — Copilot proposes, you approve.

📊

### Numbers and Slides, Made Easier

Ask Copilot in Excel a plain-language question — "which region grew fastest this quarter?" — and it proposes the analysis and explains what it found, no formula knowledge required. Ask Copilot in PowerPoint to turn that finding, or a whole Word document, into a polished slide deck in minutes.

**Exam Intel**Excel Copilot: analyse data, suggest formulas, explain trends in plain language — always verify against source data. PowerPoint Copilot: build a deck from a prompt or convert an existing document into slides.

📓

### Keeping Your Best Work

A notebook is how you tell Copilot "work only from these sources" — a set of documents, emails, and links you’ve chosen, keeping answers focused instead of wandering. When a chat response turns out to be genuinely useful, turn it into a Page: a shareable, editable document you and your team can keep refining together.

**Exam Intel**Notebooks = defined set of sources that ground Copilot’s answers. Pages = shareable, editable Copilot-generated documents for ongoing collaboration. Both sit above a single chat — they organise and preserve your work.

Section 3 / Acronym Memory

## Mnemonic Wall

Memorable acronyms and phrases to anchor key exam concepts in your memory.

✍️

GCES

**G**oal, **C**ontext, **E**xpectations, **S**ource

The four ingredients of an effective prompt. Expect questions describing a weak prompt — identify which ingredient is missing.

🤖

R & A

**R**esearcher agent & **A**nalyst agent

The two named Copilot agents. Researcher = deep research with citations. Analyst = data exploration and trend explanation.

🛡️

SAFE

**S**ources verified, **A**void sensitive data, **F**acts checked, **E**ffort stays accountable to you

Responsible AI habits for everyday business use — check citations, never paste confidential data into a prompt, fact-check the output, and remember you’re accountable for what you send.

📓

N + P

**N**otebooks organise sources, **P**ages save the output

Notebooks define what Copilot is allowed to use as its source material. Pages preserve and let you keep collaborating on a good result.

💻

WOTPE

"**W**ise **O**wls **T**ype **P**erfect **E**mails" — **W**ord, **O**utlook, **T**eams, **P**owerPoint, **E**xcel

The five core Microsoft 365 apps AB-730 expects you to know Copilot within. Each app has its own Copilot capabilities — know what each one is best at.

Section 4 / Contrast Memory

## Versus Arena

Side-by-side comparisons to sharpen your understanding of similar concepts.

vs

Copilot ChatvsCopilot in an App

Click to compare

#### Copilot Chat vs Copilot in an App

| Aspect | Copilot Chat | Copilot in an App |
| --- | --- | --- |
| Where you use it | Standalone assistant (browser, Teams, mobile) | Inside Word, Excel, PowerPoint, or Outlook |
| Best for | Open-ended questions, brainstorming | A task tied to the document or sheet you have open |
| Data grounding | Web knowledge, plus your work data if you ask | The specific file or email you’re working in |
| Starting point | A blank chat window | Whatever’s already on your screen |
| Example | "Draft a project update email" from scratch | "Summarise this document" while it’s open |

Click to flip back

vs

PromptsvsAgents

Click to compare

#### Prompts vs Agents

| Aspect | Prompts | Agents |
| --- | --- | --- |
| What it is | A single instruction you type | A multi-step assistant that plans and investigates |
| Effort | One question, one answer | Works through several steps before responding |
| Best for | Quick tasks — draft, summarise, rewrite | Complex research or data analysis |
| Example | "Rewrite this in a more formal tone" | Researcher agent producing a cited report |
| Output | A direct response | A reasoned report or analysis, with explanation |

Click to flip back

vs

Researcher AgentvsAnalyst Agent

Click to compare

#### Researcher Agent vs Analyst Agent

| Aspect | Researcher Agent | Analyst Agent |
| --- | --- | --- |
| Purpose | Deep research across documents, emails, and the web | Data analysis and pattern-finding |
| Works with | Files, meetings, emails, web sources | Spreadsheets and structured datasets |
| Output | A cited report with reasoning | Trends and explanations in plain language |
| Example use | "Research our competitors’ pricing approaches" | "What’s driving the Q3 sales dip?" |
| Style | Reads and synthesises | Calculates and explains |

Click to flip back

vs

Grounding DatavsGeneral Knowledge

Click to compare

#### Grounding Data vs General Knowledge

| Aspect | Grounding Data | General Knowledge |
| --- | --- | --- |
| Source | Your organisation’s files, emails, and meetings | Public information the model was trained on |
| Accuracy for your business | Specific and current to your work | General, and can be outdated |
| Citations | Shows sources you can check | No citations to your own data |
| Best for | "What did we agree on in yesterday’s meeting?" | "Explain what generative AI means" |
| Risk | Permissions-based — still worth verifying | Can be generic, or simply wrong for your context |

Click to flip back

Section 5 / Grouping Memory

## Cheat Sheet

Organised reference grouped by exam domain — everything you need on one page.

### Understand Generative AI Fundamentals

~33%

#### What Generative AI Is

-   Creates new content — text, images, summaries — rather than retrieving existing data
-   Powered by large language models (LLMs) predicting likely next words
-   Copilot = generative AI + your Microsoft 365 work content
-   Can be confidently wrong ("hallucinate") — always fact-check

#### Copilot Chat

-   Standalone assistant, not tied to a specific document
-   Available in browser, Teams, and mobile
-   Can use general web knowledge or your work data
-   Best for open-ended questions and brainstorming

#### Responsible AI for Everyday Use

-   Review and fact-check AI output before using or sharing it
-   Never paste sensitive or confidential data into a prompt
-   Check citations when Copilot is grounded in your data
-   You’re accountable for the final content, not the AI
-   Copilot only uses data you already have permission to see

### Manage Prompts and Conversations by Using AI

~33%

#### Writing Good Prompts (GCES)

-   Goal: what you actually want
-   Context: background info Copilot needs
-   Expectations: tone, length, format
-   Source: what to base the answer on
-   Specific prompts beat vague ones

#### Copilot Agents

-   Agents handle multi-step tasks, not just single replies
-   Researcher agent: deep research with citations
-   Analyst agent: explores data, explains trends
-   Agents may ask clarifying questions before starting

#### Refining a Conversation

-   Conversations are multi-turn — Copilot remembers context
-   Give direct feedback rather than starting over
-   Start a new chat to reset context for a new topic
-   Coach Copilot like briefing a colleague

### Draft and Analyze Business Content by Using AI

~33%

#### Copilot in Word & Outlook

-   Word: draft, rewrite, and summarise documents
-   Outlook: draft replies, summarise threads, coach tone
-   Works from a blank page or existing content
-   You always review before sending

#### Copilot in Excel & PowerPoint

-   Excel: analyse data, suggest formulas, explain trends — no formula syntax required
-   PowerPoint: build a deck from a prompt or a Word document
-   Always verify generated numbers against source data

#### Notebooks & Pages

-   Notebooks: a defined set of sources for Copilot to ground answers in
-   Pages: shareable, editable Copilot-generated documents
-   Turn a good chat response into a Page to preserve it

Section 6 / Method of Loci

## The Memory Palace

Walk through themed rooms — each object anchors a concept in spatial memory.

### The Welcome Lobby

Generative AI Fundamentals — Where you get oriented

🌱

Generative AI

Creates new content by predicting likely words — powers Copilot across Microsoft 365

💬

Copilot Chat

Standalone assistant — not tied to a document. Browser, Teams, mobile

🛡️

Responsible AI

Check sources, avoid sensitive data in prompts, stay accountable for the output

⚠️

Hallucinations

AI can sound confident and still be wrong — always fact-check before you act

### The Prompt Workshop

Prompts & Conversations — Where you brief the assistant

✍️

GCES

Goal, Context, Expectations, Source — the four ingredients of a strong prompt

🔍

Researcher Agent

Deep research across files, emails, meetings, and the web — with citations

📊

Analyst Agent

Explores a dataset, tests ideas, explains trends in plain language

🔄

Refining the Chat

Multi-turn conversations — give feedback rather than starting over

### The Content Studio

Draft & Analyze Business Content — Where the work gets done

📧

Word & Outlook Copilot

Draft, rewrite, and summarise documents and emails — you review before sending

📈

Excel & PowerPoint Copilot

Analyse data and build a deck in plain language — no formulas needed

📓

Notebooks

A defined set of sources that keeps Copilot’s answers focused

📄

Pages

Shareable, editable Copilot output you and your team can keep refining

Section 7 / Pattern Recognition

## Pattern Spotter

Decision trees and trigger-answer pairs — see the pattern, know the answer.

Which Copilot Experience Do I Need?

Which Copilot Experience Do I Need?  
  ├── Open-ended question, no document open → Copilot Chat  ├── Editing or summarising a document you have open → Copilot in Word  ├── Sending or summarising an email → Copilot in Outlook  ├── Analysing spreadsheet data → Copilot in Excel  ├── Building a presentation → Copilot in PowerPoint  ├── Deep research with citations across sources → Researcher agent  └── Explaining trends in a dataset → Analyst agent

Is My Prompt Missing Something? (GCES Check)

Is My Prompt Missing Something? (GCES Check)  
  ├── Copilot doesn’t seem to know what you actually want → Add your Goal  ├── Response ignores important background information → Add Context  ├── Wrong length, tone, or format → State your Expectations  └── Copilot guesses instead of using the right material → Point it to a Source

Should I Trust This AI Output?

Should I Trust This AI Output?  
  ├── Contains facts, figures, or names you can’t verify → Check the sources and citations first  ├── About to paste confidential or personal data into a prompt → Stop — don’t include sensitive data  ├── Response looks confident but feels slightly off → AI can be confidently wrong — fact-check it  └── Ready to send or share the content → You’re accountable — review before sending

## Decision Cards

“standalone AI assistant, not tied to a document”→Copilot Chat

“summarise this document” while it’s open→Copilot in Word

“draft a reply” or “summarise this thread”→Copilot in Outlook

“which region had the highest growth” on a spreadsheet→Copilot in Excel

“turn this into slides” or “build a deck”→Copilot in PowerPoint

a cited research report across files, email, and the web→Researcher agent

“explain the trend behind this dataset”→Analyst agent

“Goal, Context, Expectations, Source”→The GCES prompt framework

Ready to certify?

## Train with practitioners, not presenters

Lucid Labs delivers Microsoft certification training led by Microsoft Certified Trainers (MCTs) and grounded in real-world project experience. We adapt every session to your team's environment, data stack, and business objectives — because the best exam prep comes from engineers who build these solutions every day.

🎯

Tailored Content

Role-based Copilot enablement built around your team’s actual daily tasks — not generic slides.

🛠️

Hands-On Practice

Practise real prompts, agents, and everyday scenarios in Word, Excel, PowerPoint, and Outlook with expert guidance.

📈

Exam + Everyday Confidence

Pass the AB-730 exam and walk away using Copilot productively and responsibly from day one.

[Talk to us about AI Business Professional (AB-730) training](https://lucidlabs.com.au/?service=training-consulting&message=I%27m%20interested%20in%20everyday%20Microsoft%20365%20Copilot%20training%20for%20my%20business%20team.#contact)

Custom training for teams & individuals — remote or on-site across Australia

![Keith Oak](https://lucidlabs.com.au/team/koak-400.jpg)

Keith Oak

Chief Technology Officer — Lucid Labs

Microsoft Solutions Partner architect specialising in Fabric, Azure Data & AI, and GitHub Enterprise. 18+ years delivering data platforms for Australian businesses — building the systems these exams test every day.

[LinkedIn ↗](https://www.linkedin.com/in/keithoak/)[lucidlabs.com.au ↗](https://lucidlabs.com.au/)Published 29-03-2026
