Free Starter Guide
Learn AI for Work for Free (Claude, ChatGPT, Copilot, Gemini)
Night school, minus the commute: learn on your own time, at your own pace, with your own work as the homework.
This guide shows you how to learn AI for work for free, using lessons you can read without creating an account. It starts with Anthropic's Claude and adds free first lessons for ChatGPT, Microsoft 365 Copilot, and Google Gemini, so you can begin with the tool your workplace already uses. You will get a realistic weekly rhythm, a starter path in order, and honest notes on what a few weeks of practice will and will not give you.
You can learn AI for work for free by starting with Nightschool AI's free lessons. For Anthropic's Claude, the Why Claude? course is free in full and readable without an account. For ChatGPT, Microsoft 365 Copilot, and Google Gemini, the first two lessons of our own starter courses are free to start. These are Nightschool AI courses, not courses made by those companies. Begin with the tool your workplace already uses, then try one idea on a real task.
Published
Free lesson · no account neededMeet ClaudeWhy Claude?6 minStart the free lessonKey Takeaways
- The Why Claude? course is free in full and readable without an account, so you can start learning Claude AI for free today.
- The ChatGPT, Microsoft 365 Copilot, and Google Gemini starter courses are free to start: the first two lessons of each are open to read.
- Start with the tool your workplace already uses, because what you learn applies to your job the next morning.
- Two or three short sessions a week, each ending with one real task, beats a single long weekend of reading.
- Free lessons give you a solid start and good habits. They do not replace practice, and you still check AI output before anyone else sees it.
Where each free starting point begins
| Tool you want to learn | What is free | Your first lesson | A good fit if |
|---|---|---|---|
| Anthropic's Claude | The Why Claude? course, free in full | Meet Claude | You want habits that carry across tools, or your team already uses Claude |
| OpenAI's ChatGPT | The first two lessons of Your first week with ChatGPT at work, free to start | Getting an answer you can actually use | Your workplace already uses ChatGPT |
| Microsoft 365 Copilot | The first two lessons of Your first week with Microsoft 365 Copilot at work, free to start | Clear your morning inbox before your next meeting | Your workplace runs on Microsoft 365 |
| Google Gemini | The first two lessons of Your First Week With Google Gemini at Work, free to start | Clear a Crowded Inbox With Gemini in Gmail | Your workplace runs on Google Workspace |
Copy and Paste
Prompts You Can Use Today
Turn a long email thread into a decision summary
You are helping me prepare for a meeting. Below is an email thread I was copied on. Summarize it in five bullet points, list any decisions already made, list open questions that still need an owner, and flag anything that sounds like a deadline. Keep the tone neutral. Thread: [paste a thread you are allowed to share]
Swap in a thread that your workplace's AI policy allows you to paste, or an invented one while you practice.
Draft a reply, then tighten it
Draft a reply to the message below from my manager. I want to agree to the new deadline but ask for one extra day for review. Write it in four sentences or fewer, in a friendly but direct tone. Then give me a second version that is even shorter. Message: [paste the message]
Change the request, the length, and the tone to match your real situation. Read both versions aloud before sending either.
Get a plain-English explanation of something you must understand this week
I work in [your role] and I have a meeting about [topic] on Thursday. Explain it to me as if I am smart but new to it. Use one worked example from a workplace like mine, define any jargon the first time it appears, and end with three questions I could ask to sound prepared.
Replace the role and topic. Afterwards, check the key facts against a source you trust.
Make the assistant find the weak points in its own answer
Look at the answer you just gave me. List every factual claim, number, or name in it that I should verify before I rely on it, and say how confident you are in each. Then point out one assumption you made about my situation that I never told you.
Use this after any answer you plan to share with a colleague or client. It does not replace your own check.
Build your own two-week practice plan
I work full time as a [your role] and can spare about three short sessions a week to learn how to use an AI assistant on my job. Suggest a two-week plan. Each session should name one real task from my work, one thing to try, and one way to check the result. My main tools are [list them].
List the software you really use, such as your email, documents, and spreadsheets, so the tasks are concrete.
The Free Starter Path
Start with a free lesson
These four lessons can be read without an account, in this order. Begin with the first, then continue with the one for the tool your workplace uses.
- 1. Why Claude?Meet Claude6 min
- 2. Your first week with ChatGPT at workGetting an answer you can actually use14 min
- 3. Your first week with Microsoft 365 Copilot at workClear your morning inbox before your next meeting11 min
- 4. Your First Week With Google Gemini at WorkClear a Crowded Inbox With Gemini in Gmail12 min
The Free Options
Four free ways to start with free AI lessons for work
You do not need a budget to begin. Each option below can be read without an account, and each is built for people who have a job to do.
Claude, free in full
The Why Claude? course has six lessons, from Meet Claude to Make Claude Sound Like You. It is free in full, and it is the quickest way to learn how to talk to an AI assistant well.
ChatGPT, free to start
The first two lessons of Your first week with ChatGPT at work are open to read. They are part of a six-lesson Nightschool AI course, which is free to start rather than free in full.
Microsoft 365 Copilot, free to start
The first two lessons of Your first week with Microsoft 365 Copilot at work are open to read. The full course has six lessons, and it is made by Nightschool AI, not by Microsoft.
Google Gemini, free to start
The first two lessons of Your First Week With Google Gemini at Work are open to read. The course is ours, it has six lessons, and the first two are open to everyone.
Read without an account
The free lessons open straight away, with no sign-up wall between you and the first page. You can decide whether the teaching suits you before you give anyone your email.
Choosing a Tool
Start with the tool your workplace already uses
The most useful first tool is usually the one your employer already provides. Your colleagues use it, someone in IT has probably written rules for it, and whatever you learn can be applied to a real task the next morning. Learning a tool you cannot open at your desk is a fine hobby, but it is a slow way to get better at your job.
Before you paste anything real into any AI assistant, find out what your employer allows. Many workplaces have a policy on which tools are approved and what kind of information can go into them. If you cannot find one, ask your manager or your IT team, and until you have an answer, practice with public material or invented examples. This habit matters more than any prompting trick.
If your workplace has no preference, or you are choosing for yourself, begin with the Why Claude? course. It teaches habits that carry across every assistant: giving context, stating what a good result looks like, iterating on a draft, and checking the output. Those habits are the real skill, and menus and button names change far more often than they do.
Once you have the habits, move to the lessons for your own setup. If your files and calendar live in Microsoft 365, the Copilot lessons are the natural next step. If your team works in Gmail and Google Docs, take the Gemini lessons. If colleagues already share ChatGPT chats, begin there. You do not need to learn all four. Most people are well served by one tool they use well.
A Realistic Plan
A weekly rhythm for free AI training for work, alongside your job
A full-time job leaves small pockets of time, not open afternoons, so the plan has to fit the pockets. Aim for two or three short sessions a week, around two or three hours in total, and spread them out. A lunch break, a commute, or an evening after dinner all work. Short, regular sessions beat one long weekend of reading because you need time to try an idea on your own work and notice what happened.
Make the first session a reading session only. Open a lesson, read it at your own pace, and write down one task from your actual week that the lesson makes you think of. Do not try to practice yet. The goal is to leave with a single concrete idea, such as summarizing a long thread or drafting a first version of a status update.
Make the second session a practice session. Take that one task and try it with the assistant, using whichever tool your workplace allows. Compare what you get to what you would have written yourself, and change one thing at a time: add context, name the audience, ask for a shorter version. This is where learning AI for work for free turns into a skill you can feel.
Use the third session, if you have one, to review. Look back at what worked, save the prompts you would reuse, and read the next lesson. Over a few weeks, you will build a small personal library of prompts for your own tasks. That library is worth more than any amount of passive reading, and it is the only part of the process that is entirely yours.
The Starter Path
The starter path, lesson by lesson
The path below runs in a set order, and every lesson on it can be read without an account. If you only have time for one tool, do the first lesson, then jump to the lesson for the tool your workplace uses. Each step ends with something to try on your own work during the same week.
Step one is Meet Claude, the first lesson of the Why Claude? course. It is about six minutes long and introduces Anthropic's Claude in plain terms. This week, take one email or document you already understand well and ask Claude to summarize it. Then check the summary against what you know, which teaches you how reliable the answer really is.
Step two is Getting an answer you can actually use, the first lesson of Your first week with ChatGPT at work. It runs about fourteen minutes and focuses on turning a vague request into a useful result. This week, ask a work question twice: once as you would normally, and once after adding who the answer is for and what format you need. The second lesson, Keeping a project's chats and files together (about eleven minutes), is also free and helps once you have more than one conversation going.
Step three is Clear your morning inbox before your next meeting, the first lesson of Your first week with Microsoft 365 Copilot at work. It takes about eleven minutes. This week, pick a morning when your inbox is busy and apply the lesson's approach before your first meeting. The second free lesson, Draft and tighten a document without starting from a blank page (about ten minutes), is the one to take when you next have a report or proposal to write.
Step four is Clear a Crowded Inbox With Gemini in Gmail, the first lesson of Your First Week With Google Gemini at Work. It runs about twelve minutes. This week, use it on a day when your mail has piled up, and note how much time you save and what you had to correct. The second free lesson, Write a First Draft in Google Docs With Gemini (about eleven minutes), follows naturally when you have a document to start.
Practice on Real Work
Five safe ways to practice on your own job
Reading a lesson is the easy half. These habits turn the lessons into skill, and none of them requires sharing anything your employer would not want shared.
Summarize something you already know
Give the assistant a document you understand well, then compare its summary with your own. You learn where it helps and where it drifts without risking a decision.
Draft, then rewrite in your voice
Ask for a first draft of a routine message, then edit it until it sounds like you. Notice which edits you make every time, and put them into your next prompt.
Prepare for a meeting
Ask for a list of questions you might be asked and a list of questions you should ask. Treat the output as a prompt for your own thinking, not a script.
Ask it to find the holes
After you finish a piece of work, ask the assistant what is missing, unclear, or unsupported. You decide which points matter and which to ignore.
Keep a small prompt log
Save each prompt that worked, plus one line on what you changed. After a few weeks, you will have a personal toolkit for the tasks you repeat.
Learning Claude
How to learn Claude AI for free, step by step
If your question is how to learn Claude AI for free, the answer starts with the Why Claude? course. It has six lessons: Meet Claude, Your First Conversation, Context Windows & System Prompts, Claude's Strengths & Limits, From Chatting to Building, and Make Claude Sound Like You. The whole course is free and can be read without an account, so you can go through it at your own pace.
Read the lessons with Claude open in another window, and pause after each one to try what it describes. Your First Conversation is a good place to stop and spend an evening on a real task. Claude's Strengths & Limits is worth reading slowly, because it tells you where to trust the output and where to check it, which saves you from the most common early mistakes.
After Why Claude?, two more courses are free in full: The Three Modes, with five lessons that explain the different ways to use Claude and how to choose between them, and Master the Essentials, with four lessons on iterating, analyzing documents, researching, and making decisions. Strategic Prompting also has free lessons, including Context Is Everything and Claude Prompting Fundamentals for Everyday Work. Together these give you a solid base without paying for anything.
To use Claude itself, visit Anthropic's own site and see how access works for you or your workplace. Nightschool AI is a separate, independent learning site, so it is the place for the lessons, and Anthropic is the place for the product. Keeping that distinction clear will save you confusion later.
Honest Limits
What a few weeks of free practice will and will not give you
A few weeks of regular practice will give you real, usable skills. You will write better requests, spot weak answers sooner, and know which of your routine tasks an assistant helps with and which it does not. You will also feel less intimidated by the tools, which matters more than it sounds, because comfort is what makes people keep practicing.
It will not make you an expert, and the free lessons are a start rather than a full education. The ChatGPT, Copilot, and Gemini courses are free to start, meaning their first two lessons are open and the rest of each course goes beyond the free portion. Treat the free lessons as a way to find out whether the teaching fits you, and as a genuine foundation in their own right.
The most important limit is that AI output can be wrong in ways that sound confident. Assistants can invent a figure, misread a document, or miss the point of what you asked. Before anything reaches a colleague or a client, read it yourself, verify names, numbers, and dates against a trustworthy source, and make sure it says what you mean. You remain responsible for the work, and no tool changes that.
A second limit is confidentiality. Customer details, unreleased plans, and personal data of colleagues are not material to paste into a tool your employer has not approved. When in doubt, practice with invented or public material, and ask your manager or IT team what is allowed. Being careful here is part of being good at using AI at work, not an obstacle to it.
Judging Quality
How to tell whether a free AI course is actually good
Free does not tell you much about quality. These five checks work on any course, ours included, and take a few minutes to apply.
It makes you do something
A good lesson ends with a task you can try on your own work. If you finish a course without having typed a single prompt, you have read about AI rather than learned it.
It names specific work
Look for lessons about inboxes, documents, and meetings, not vague promises. Specific tasks are easier to try and easier to judge.
It is honest about limits
A trustworthy course tells you where the tool fails and how to check its output. Be wary of any lesson that only describes what the assistant gets right.
It respects your time
Short lessons that you can finish in one sitting fit around a job. A course you can read in pieces and return to is easier to keep up than one that demands blocks of hours.
It lets you look before you commit
If you can read the first lessons without an account, you can judge the teaching on its merits. That is a reasonable test for any free offering.
Is It Worth It?
Free, no degree required, and still worth your evenings
You do not need a degree, a technical background, or any coding to learn how to use AI assistants at work. The skills involved are mostly the ones you already use: describing what you want clearly, giving relevant background, judging whether a result is good, and revising. People in every kind of role, from administration to sales to operations, can learn these without special preparation.
Whether free AI courses are worth your time depends on what you do afterwards. A course you read and forget is worth little, whatever it costs. A course that sends you back to your own tasks with one new technique each week is worth a lot, because you will keep using what you learned.
The practical test is simple. After your first week, ask whether you have a task you now do differently, or a prompt you would reuse. If yes, keep going with the next lesson. If not, change one thing, perhaps the tool, the task, or the time of day you practice, and try again before deciding the approach does not work for you.
Read the first lesson, then try it on your own work
Reading about AI and being good at it on your own job are different things. The Nightschool AI lessons are hands-on from the first one, and you can read the free ones without an account. Pick the lesson for the tool you use, and try one idea on a real task this week.
Frequently Asked Questions
How do I learn Claude AI for free?
Start with the Why Claude? course on Nightschool AI. It has six lessons, from Meet Claude to Make Claude Sound Like You, and it is free in full and readable without an account. Read each lesson with Claude open beside it, and try one idea on a real task after every lesson. Then continue with The Three Modes and Master the Essentials, which are also free in full.
Can I learn AI for work for free?
Yes, you can learn AI for work for free, at least to a solid starting level. Nightschool AI's Why Claude? course is free in full, and the ChatGPT, Microsoft 365 Copilot, and Google Gemini starter courses each have two free lessons. Pair them with regular practice on your own tasks, and you can build real habits without paying for anything. Mastery takes longer than a few weeks, but a useful start does not.
Are there free lessons for ChatGPT, Copilot, or Gemini?
Yes. Nightschool AI has its own starter courses for ChatGPT, Microsoft 365 Copilot, and Google Gemini at work, and the first two lessons of each are free to read. They are free to start, not free in full. These courses are made by Nightschool AI, not by OpenAI, Microsoft, or Google, and they teach the tools from a working person's point of view.
Do I need an account to read the free lessons?
No. The free lessons can be read without an account, so you can open the first page and judge the teaching before you sign up. An account is optional, and you only need one if you want the site to remember your progress or to go beyond the free portion of a course.
Which AI tool should I learn first for work?
Learn the tool your workplace already uses and has approved. Your colleagues will use it too, and you can apply each lesson the next morning. If you have no preference, begin with the Why Claude? course, because the habits it teaches, such as giving context, iterating, and checking output, carry over to every other assistant. Then take the lessons for your own setup.
Is Nightschool AI affiliated with Anthropic, OpenAI, Microsoft, or Google?
No. Nightschool AI is independent and is not affiliated with, endorsed by, partnered with, or sponsored by Anthropic, OpenAI, Microsoft, or Google. Claude is Anthropic's product, ChatGPT is OpenAI's, Microsoft 365 Copilot is Microsoft's, and Gemini is Google's. We name them only to describe the tools our own lessons teach.
Which free AI course is actually good?
A good free course makes you practice on real work, names specific tasks, and is honest about where the tool fails. It should also let you read the first lessons before committing. Judge any course, including ours, by whether you do something differently after the first week. The checklist earlier on this page gives five quick tests you can apply to any course.
Are free AI courses worth it?
They are worth it if you use them. A free course you read and forget has little value, while one that gives you a technique to try on your job each week builds a useful skill. Free is a good way to find out whether the teaching style suits you. If it does, you will know it within a few lessons, and if not, you have lost little.
Can I learn AI for free without a degree?
Yes. Using AI assistants at work does not require a degree, coding, or a technical background. The core skills are describing what you want clearly, supplying context, judging the result, and revising, and most working people already practice these. The free lessons here are written for people with jobs, not specialists, and you can go through them at your own pace.
How long does it take to get comfortable with AI at work?
With two or three short sessions a week, many people feel comfortable with the basics within a few weeks. That means you can write a clear request, spot a weak answer, and know which routine tasks the tool helps with. Being truly skilled takes longer and comes from repeated practice on your own tasks, not from reading more.