What Is AI? A Complete Overview of AI Basics Every Office Worker Should Know
A hands-on guide to AI fundamentals for office workers and B2B sales reps. By the end of this guide, you will be able to speak up with confidence in AI-related meetings and apply AI tools to your daily work right away.
Before you begin: how to use this guide
Time needed: about 60–90 minutes (reading plus hands-on practice)
What you need: a PC or laptop with an internet connection and a web browser (Chrome recommended)
Prior knowledge: None. This guide is written for complete beginners with no technical background
Goal: understand 7 key AI terms and write your own prompts in ChatGPT
Before/After Comparison
Category | Before | After |
|---|---|---|
Understanding of AI terms | “Machine learning? Deep learning? Aren’t they all the same thing?” | Can tell the 7 key terms apart and explain them |
Experience with AI tools | Have never used one | Have written 3+ work-related prompts in ChatGPT |
Meeting participation | Stays silent when AI comes up | Suggests, “Generative AI can solve that.” |
Work efficiency | 2 hours to draft a report | First draft done in under 30 minutes (about 75% time saved) |
Step 1. Understand the definition of artificial intelligence (AI) in your own words
1-1. What is AI?
**Artificial intelligence (AI)** is, in a word, *“technology that gives computers the ability to think and make judgments like a person.”*
Let’s look at some examples of AI you already use every day:
Smartphone face recognition: the camera tells your face apart from others → image recognition AI
Naver Smart Lens: take a photo and it searches for the product → visual AI
YouTube recommendations: picks videos that match your taste → recommendation AI
Kakao Translate: automatically converts foreign languages into Korean → natural language processing AI
💡 Pro tip: Don’t overthink AI. The smartphone you use every day already has 10+ AI features on average. AI is already part of your daily life.
1-2. Hands-on: ask AI what AI is
Let’s go to ChatGPT right now and ask the AI directly, “What is AI?”
How to do it:
Open a web browser and type
chat.openai.comin the address barClick the Sign up button (if you don’t have an account)
Enter your email address and set a password (you can also sign up quickly with a Google account)
Once you’re logged in, you’ll see a text input box at the bottom of the screen
Copy the prompt below exactly and paste it into the input box
Click the arrow button (→) on the right of the input box, or press Enter
I’m an office worker with no IT background. Please explain what artificial intelligence (AI) is using an analogy that even an elementary school student could understand. Keep your answer under 300 characters.
Expected result: The AI will explain it with analogies like “a clever parrot” and “a blazing-fast calculator.”
💡 Pro tip: Add the condition *“so even an elementary school student can understand”* to your prompt, and the AI will answer in plain language instead of jargon. This is the foundation of prompt engineering.
Step 2. Understand the 3 Core Categories of AI
AI has several sub-concepts. The easiest way to understand them is to think of **Russian nesting dolls (matryoshka)**.
2-1. Understanding from the largest doll to the smallest
🔵 Artificial Intelligence (AI) ← The broadest concept
└─ 🟢 Machine Learning ← A field within AI
└─ 🟡 Deep Learning ← A method within machine learning
└─ 🔴 Generative AI ← An application built on deep learning
Concept | Simple analogy | Everyday example |
|---|---|---|
Artificial Intelligence (AI) | Every computer that makes human-like judgments | Includes self-driving cars and voice assistants |
Machine Learning (ML) | Learning by studying lots of data and finding patterns on its own | Spam email filtering |
Deep Learning (DL) | Advanced learning that mimics the brain’s neural networks | Telling cats from dogs in photos |
Generative AI | AI that creates something new based on what it has learned | ChatGPT, Midjourney, CLOVA X |
2-2. Fully understanding it through an analogy
“AI is the big idea of ‘cooking.’ Machine learning is ‘figuring out recipes on your own.’ Deep learning is ‘a Michelin-level method for developing refined recipes.’ And generative AI is ‘the chef who uses those recipes to actually create new dishes.’”
Remember this analogy, and you can confidently answer when someone in a meeting asks about the difference between “machine learning” and “deep learning.”
2-3. Hands-on: Check the differences with AI
Enter the following prompt in the ChatGPT input box:
Please summarize the differences between AI, machine learning, deep learning, and generative AI in a table. For each one, include a one-line analogy and a representative service name. Write it for readers with no technical background.
How to do it:
Paste the prompt above into the ChatGPT input box
Press Enter
When the answer appears as a table, click the **copy icon (📋)** at the top right of the screen
Paste it into Notepad or Notion and save it
💡 Pro tip: Don’t take AI answers at face value. Make it a habit to cross-check against 2–3 sources. AI can sometimes produce information that sounds plausible but is wrong (hallucinations).
Step 3. 7 AI Terms Every Professional Should Know
3-1. Glossary (bookmark this)
No. | Term | Pronunciation | Meaning | Examples for Office Workers |
|---|---|---|---|---|
1 | LLM | LLM | Large language model. AI trained on vast amounts of text | The core technology behind ChatGPT, Claude, and Gemini |
2 | Prompt | Prompt | A question or instruction you give to AI | “Draft this quarter’s sales report.” |
3 | Token | Token | The smallest unit of text that AI processes | One Korean character ≈ 1–2 tokens; cost depends on token count |
4 | Hallucination | Hallucination | When AI confidently gives information that isn’t true | “Is this number right?” Always fact-check |
5 | Fine-tuning | Fine-tuning | Further training a base AI model for a specific domain | Retraining AI on your company’s sales data |
6 | RAG | RAG | A technique that gives AI external documents to reference, improving accuracy | Connect an internal manual PDF to AI and ask it questions |
7 | Multimodal | Multimodal | AI that processes multiple formats at once, such as text, images, and voice | Upload a photo and the AI analyzes its content |
3-2. Practice: Writing Prompts with These Terms
Enter the prompt below into ChatGPT:
I’m a B2B sales rep. At next week’s client meeting, I need to explain “how we’ve applied AI to our product.” Write a 3-minute explanation script that naturally includes the terms LLM, prompt, and RAG and that a non-technical client executive can understand.
How to use the result:
Copy the response
Paste it into Word or Google Docs
Edit it with your own product name and specific features
Practice by reading it aloud three times
Before/After Comparison: Meeting Prep
Category | Before | After |
|---|---|---|
Script prep time | 3 hours (research + writing) | 20 minutes (AI draft + edits) |
Accuracy of terminology | “AI handles it for us” (vague) | “We used RAG to train on client documents and improved accuracy” (specific) |
Customer trust | “They don’t seem to know their stuff...” | “They really know their stuff.” |
Step 4. Try Generative AI Tools Hands-On
4-1. Comparing Popular Generative AI Tools
Tool | Developer | Free? | Strengths | Best for |
|---|---|---|---|---|
ChatGPT | OpenAI | Free (basic) / Paid ($20/month) | Versatility, long conversations | Reports, emails, brainstorming |
Claude | Anthropic | Free (basic) / Paid ($20/month) | Long-document analysis, safety | Contract review, summarizing long reports |
Gemini | Free (basic) | Google services integration | Google Workspace users | |
Copilot | Microsoft | Free (basic) | Microsoft Office integration | Excel and PowerPoint work |
Wrtn | Wrtn Technologies | Free | Optimized for Korean | Writing Korean-language content |
4-2. Practice: Writing a Business Email with ChatGPT
Scenario: You need to write an email proposing a meeting time to a client.
Step-by-step instructions:
Go to
chat.openai.comand log inClick the “New chat” button at the top left to start a new conversation
Enter the prompt below:
I’m a B2B sales rep at an IT solutions company. Write a business email to my client, Mr. Kim, proposing a product demo meeting next Tuesday or Wednesday afternoon. The tone should be polite but friendly, around 100 words, and include a subject line. Our product is called ‘SmartFlow AI’, and its key feature is automated sales pipeline analysis.
Press Enter and review the result
If anything isn’t quite right, type a follow-up request in the input box:
“Please make the tone more formal”
“Add a recent customer example as a P.S.”
“Please write a Korean version as well”
Copy the final result and paste it into your email program
💡 Practical tip: Don’t expect a perfect result on the first try. Working with AI is like a game of ping-pong. Get a first answer → request edits → get a new version, and after 2–3 rounds you’ll have a result you’re happy with. This is called **“Iterative Refinement.”**
Before/After comparison: Writing an email
Category | Before (written by hand) | After (with AI) |
|---|---|---|
Turnaround | 15–20 minutes | 3–5 min |
Typos | Occasionally | Rarely |
Tone consistency | Varies with each draft | Consistent business tone every time |
Multilingual support | Requires a separate translation tool | Write in English and Korean at once |
Step 5. How to Write Good Prompts (Prompt Engineering Basics)
5-1. The 5 Elements of a Good Prompt — The RCTSO Formula
Element | English | Meaning | Example |
|---|---|---|---|
R - Role | Role | The expert role you give the AI | "You are a B2B sales consultant with 10 years of experience." |
C - Context | Context | Situation and background information | "I have a presentation for a large company’s CTO next week." |
T - Task | Task | The specific job to be done | "Write a 15-minute presentation script." |
S - Style | Style | The tone and format you want | "Professional but in plain language, with analogies." |
O - Output format | Output | Specify the shape of the result | "As a Markdown table, split slide by slide." |
5-2. Practice: Bad Prompt vs. Good Prompt
Bad prompt (vague, lacking information):
Make me sales materials
Good prompt (RCTSO formula applied):
[Role] You are a B2B sales strategist with 10 years of experience in the SaaS industry. [Context] Our company sells a cloud-based HR management solution called 'PeopleFlow'. Next week I have a first meeting with the HR director of a 500-employee manufacturer. [Task] Write the text for a one-page product brief to use in this meeting. [Style] Use a professional yet easy-to-understand tone, and emphasize three points that differentiate us from competitors. [Output format] Structure it in this order: title, subtitle, body (3 sections), and CTA (call to action).
Try it yourself:
In ChatGPT, enter the bad prompt first and review the result
Open a new chat, enter the good prompt, and review the result
Compare the specificity, usefulness, and completeness of the two results for yourself
💡 Pro tip: The more specific the prompt, the better the result. According to a McKinsey report, employees who write good prompts are about 40% more productive when using AI than those who don’t.
Step 6. 5 AI Work Scenarios You Can Use Right Away
6-1. Prompt Templates by Scenario
Scenario | Tools Used | Estimated Time Saved |
|---|---|---|
① Meeting Minutes Summary | ChatGPT / Claude | Before: 30 min → 5 min |
② Competitor Analysis Report | ChatGPT + Web Search | Before: 4 hours → 1 hour |
③ Sales Proposal Draft | ChatGPT / Copilot | Before: 3 hours → 40 min |
④ Excel Formulas / Macros | ChatGPT / Copilot | Before: 1 hour → 10 min |
⑤ Customer FAQ Response Scripts | ChatGPT / Claude | Before: 2 hours → 20 min |
6-2. Hands-On: Automating Meeting Minutes
Scenario: You need to share the key points of a one-hour meeting with a colleague.
How to do it:
Prepare the notes you took during the meeting as text (a list of keywords is fine)
Enter the following into ChatGPT:
Below are the key keywords and notes from today’s weekly marketing team meeting. Please turn them into meeting minutes with four sections: (1) a 3-line summary, (2) key discussion points (itemized), (3) decisions made, and (4) next action items (with owners). Keep the tone concise, suitable for sharing internally.
[Notes]
Social media ad performance down 15% from last month
Kim (Associate): Target age group needs to be redefined
Park (Manager): Proposes reallocating budget (Instagram → YouTube)
Revised plan due next Friday
Brainstorming session for this month’s new campaign is on Wednesday
Press Enter and review the result
Paste the result into your team messenger (Slack, Teams, etc.) to share it
💡 Pro tip: Summarizing meetings with AI reduces missed items and keeps the format consistent every time. Team members are happier with it, and the minutes are far easier to search later.
Step 7. 3 Things to Watch Out for When Using AI
7-1. Security and Privacy
Never enter: customer personal information (names, contact details, national ID numbers), confidential company sales data, full contract text, passwords
Safe to enter: general work questions, analysis requests based on public information, hypothetical scenarios
💡 Pro tip: First check whether your company has an AI Usage Policy. If not, ask your IT department or legal team to clearly define what is permitted. As of 2024, about **62%** of major Korean corporations have established internal AI usage guidelines.
7-2. Handling Hallucinations (When AI Gets It Wrong)
AI answers confidently even when it doesn’t know. This is called *hallucination*.
3 steps to handle it:
Be skeptical: Always doubt answers that include figures, quotations, or proper nouns
Verify: Cross-check with a Google search, the original documents, or an expert
Ask for sources: Follow up with “Please tell me the source for the above” (note that the source itself also needs to be verified)
7-3. AI is a tool, not a decision-maker
What AI does well | What AI can’t do |
|---|---|
Drafting | Final judgment and accountability |
Analyzing data patterns | Ethical decision-making |
Brainstorming ideas | Emotional connection with customers |
Automating repetitive tasks | Creative strategy |
Multilingual translation | Reading cultural nuance |
Think of AI as a **"very capable intern."** It’s fast and diligent, but a person must always do the final review and make the final call.
Step 8. Next steps: your AI routine, starting today
8-1. Your first-week action plan
Day | Action items | Turnaround |
|---|---|---|
Monday | Sign up for ChatGPT + complete Steps 1–3 of this guide | 30 min |
Tuesday | Step 4 email-writing practice + apply it to 1 real work task | 20 min |
Wednesday | Step 5 practice with the RCTSO prompt formula | 20 min |
Thursday | Step 6 meeting-minutes summary practice | 15 min |
Friday | Summarize the week’s AI results + share them with a colleague | 15 min |
8-2. Your growth roadmap
[Week 1] Learn core AI concepts + first use of ChatGPT ← You are here
[Weeks 2–3] Use AI on at least 1 work task every day
[Week 4] Build your own library of 10+ prompt templates
[Month 2] Present your AI use cases to your team
[Month 3] Build AI automation workflows (integrate with Zapier, Make, etc.)
8-3. Recommended learning resources (all free)
Official ChatGPT guide:
help.openai.comGoogle AI basics course: Google’s "AI Essentials" (Coursera, Korean subtitles available)
Naver Boostcourse: free "AI for Everyone" course
YouTube: the AI beginner series on the "Nomad Coders" channel
Wrap-Up Checklist
Check what you learned today:
I can explain the difference between AI, machine learning, deep learning, and generative AI in one sentence each
I signed up for ChatGPT and entered at least 3 prompts
I remember the 5 elements of the RCTSO prompt formula
I know what hallucination is and how to deal with it
I know which types of information must never be entered into AI
I have chosen at least 1 scenario for applying AI to my real work starting next week
If you checked 4 or more of the 6, you are already in the top **30%** of AI users in the workplace. According to the 2024 Microsoft Work Trend Index, 75% of office workers use AI at work, but fewer than 25% understand and use it systematically.
💡 Final tip: AI isn’t perfect, but the productivity gap between people who use AI and people who don’t is already widening. By reading this guide to the end and practicing along, you’ve already taken an important first step. Tomorrow at work, start with just one thing: write one email with AI.