AI Training for Work, Whichever AI Tool You Use
Practical skills for employees and working professionals
AI training for work teaches you to hand an AI assistant a real task, give it the context it needs, and check what comes back before you use it. Those skills stay the same whether your company gave you Anthropic's Claude, Microsoft 365 Copilot, OpenAI's ChatGPT or Google Gemini. This page covers what good training includes, how long it takes, and how to pick a course for your tool.
Free to start, no card required. The free first Claude lesson takes about 6 minutes.
AI training is short, hands-on practice in using an AI assistant for real work: writing clear requests, giving the right context, choosing tasks that suit the tool, and checking the output. For most office jobs it needs no coding and no math. Our free Claude starter path is 15 lessons and about 97 minutes in total, which fits in a week at 20 minutes a day, and the habits carry over from one AI tool to the next.
At a glance
Nightschool AI training at a glance
- Cost
- Free to start, with no card. Full Access is a one-time unlock that opens every course; the price is shown at checkout.
- Level
- Starts at beginner; no coding needed for the starter courses.
- Format
- Self-paced lessons where you do each task yourself.
- Time
- About 97 minutes for the free Claude starter path, or 20 minutes a day for 5 days.
- Tools
- Starter courses for Claude, Microsoft 365 Copilot, ChatGPT, Gemini and Meta Muse.
- Languages
- English and 13 other languages for the Claude, Copilot, ChatGPT and Gemini starters.
- Certificate
- Yes. Finish a course and you get a certificate of completion with a link anyone can verify, ready to add to LinkedIn.
What it covers
What AI training for work should cover
Learning an AI assistant for work is closer to learning a spreadsheet than learning to program. A spreadsheet has a few moves that do most of the work, like formulas, sorting and charts, and you get good at them by using them on your own numbers. AI assistants work the same way: a handful of habits do most of the work, and you build them on your own tasks.
Giving context
An AI assistant only knows what you tell it and what it can see. You learn to include the goal, the audience, the source material and the format you want, so the first draft is close to usable instead of generic.
Picking the right tasks
Some work suits an assistant well: first drafts, summaries of long documents, reformatting, listing options, explaining a spreadsheet formula. Some work does not, like final figures nobody has checked or decisions that need your judgment. Training should teach you the difference.
Working in rounds
Good results usually come from a short back and forth: a draft, a correction, a tighter draft. You learn to say what is wrong and what to keep, which is faster than starting over with a new request.
Checking the output
AI assistants can state wrong facts in a confident tone. You learn a routine for checking names, numbers, dates, quotes and sources against the original before anything leaves your hands.
Handling work data safely
You learn to find out what your company allows you to paste into an AI tool, and why the answer can differ between a personal account and a company account.
Going further when you are ready
Once the basics are steady, training can move on to reusable instructions, working with your own files and automating tasks you repeat. You don't need any of this to get value in your first week.
How long it takes
How long AI training takes
The basics take less time than most people expect, and real fluency takes longer than one video. The figures below come from the lesson lengths in our own courses.
- First useful result: the free first Claude lesson takes about 6 minutes, and you write real prompts in it.
- A working foundation: the free Claude starter path is 15 lessons and about 97 minutes in total. At 20 minutes a day, that is about 5 days.
- Fluency in your role: a few weeks of using the assistant on your own tasks, with a course that matches your tool and your job.
Short, regular sessions tend to stick better than one long one, because each lesson asks you to try the skill on a task before you move on.
What you'll be able to do
What you'll be able to do after AI training
- Turn a messy set of notes into a clear email, memo or summary, then check it against the notes.
- Ask an assistant to explain a document, a policy or a spreadsheet, and know which parts of the answer to verify.
- Get a first version of a report, proposal or slide outline, then edit it into your own voice.
- Write reusable instructions for tasks you repeat every week.
- Tell when a task is a poor fit for AI, and do that one yourself.
- Show a manager or a teammate how you checked what the assistant produced.
Checking AI output
How to check AI output before you use it
A short routine covers most of the risk, and on a typical draft it takes a few minutes. It works the same way in every AI tool.
- 1Check every name, number and date against the material you gave the assistant, or against a source you trust.
- 2Open every link and citation. If a source doesn't exist or doesn't say what the assistant claims, treat the whole answer as suspect.
- 3Read for what is missing. Assistants summarize what is there and rarely tell you what they left out.
- 4Ask the assistant where each claim came from in your material, then confirm it yourself.
- 5For anything public, legal, financial or medical, have a person who knows the subject review it before it goes out.
- 6Keep a short note of what you checked, so you can explain it if someone asks.
An honest limit
What AI training can't do
No course makes AI output safe to send without reading it. Assistants still get facts wrong, misread tables and cite sources that don't exist. Training makes you faster at catching those mistakes and clearer about which tasks to keep for yourself, and you still have to check.
A course also can't tell you your employer's rules on AI and data. Those come from your company, so look them up before you paste work material into any AI tool.
Pick a course for your tool
How to pick AI training for the tool you use
Start with the assistant you already have at work. The core habits transfer, but menus, file access and what your license includes differ from tool to tool, so a course built for your tool saves you translating every step.
Not sure where you stand with AI yet? See the four-level map
Want a prompt for a task on your desk today? Try the free AI prompt generator.
Anthropic's Claude
Claude is a general-purpose assistant that people use for writing, analysis and long documents, and, through Claude Code, for software work. Our Claude courses run from a free first lesson for beginners to building agents.
Microsoft 365 Copilot
Copilot works inside Word, Excel, PowerPoint, Outlook and Teams, depending on your Microsoft 365 plan. What you can do also varies with your organization's settings, so a good course starts by showing you how to tell what you have.
Go to the courseOpenAI's ChatGPT
ChatGPT is a standalone assistant that many teams use for drafting, analysis and answers from their own files. A company workspace and a personal account can have different features and different data rules, so check which one you are using.
Go to the courseGoogle Gemini
Gemini works inside Gmail, Docs, Sheets, Meet and Drive for people on an eligible Google Workspace or Google AI plan. Features vary by plan, and your admin controls which ones are switched on.
Go to the courseMeta Muse
Meta's Muse is a personal AI agent you talk to in the Muse app or in WhatsApp, and it takes actions on your behalf, checking with you first before sending an email or making a purchase. Meta says it is rolling out in the US, free for most of what people need, with subscription plans for more.
Go to the courseFour checks for any AI course
- It is hands-on: you do the task in each lesson, and watching is optional.
- It uses work tasks close to yours, such as email, documents, spreadsheets and meetings.
- It teaches checking the output as a habit, in every lesson.
- It was updated after the tool's most recent major change.
By role and industry
AI training for your role and industry
The habits are the same in every job, and the tasks are not. These guides show the work people in each role and industry hand to an AI assistant. They use Claude for the examples, and the same steps work in other assistants.
Training a whole team
Rolling out AI training to employees
This page is for individuals choosing AI training for their own work. If you are setting up AI training for employees across a team, with invite codes and onboarding, Nightschool for Teams covers that.
See training for teamsCommon questions
AI training: frequently asked questions
What is AI training?
AI training is practice in using AI assistants for real tasks: writing clear requests, giving context, picking suitable tasks and checking the output. For most jobs it is about using a tool well, and it needs no coding or math.
What is AI training for employees?
It is AI training aimed at everyday work, such as email, documents, spreadsheets, meetings and research, done with the assistant the company provides. Good programs use tasks from the learner's own job and point people to the company's rules on data.
How long does AI training take?
A first useful result takes minutes. Our free first Claude lesson takes about 6 minutes, and the free Claude starter path is about 97 minutes of lessons, or about 5 days at 20 minutes a day. Fluency in your role comes from a few weeks of using the assistant on your own tasks.
Do I need to know how to code?
No. Using an AI assistant at work is mostly clear writing and good judgment. Coding courses, such as the ones on Claude Code, are there later if you want them.
Which AI tool should I train on?
The one your employer provides, or the one you already use. The core habits carry over between tools, but menus, file access and data rules differ, so a course built for your tool is easier to apply.
Is Nightschool AI training free?
Nightschool AI is free to start, with no credit card. The starter courses are free, and Full Access unlocks every course.
Does Nightschool AI give a certificate?
Yes. When you finish every lesson and pass every exercise in a Nightschool AI course, you get a certificate of completion with the name you choose, the course and the date, plus a public link anyone can use to check it. You can add it to your LinkedIn profile. It is a certificate of completion for a Nightschool AI course, not a certification, and no AI vendor issues or endorses it. Some courses help you prepare for official certifications offered by AI vendors, and the exam and the credential come from the vendor.
Is Nightschool AI affiliated with Anthropic, Microsoft, OpenAI, Google or Meta?
No. Nightschool AI is independent and is not affiliated with or endorsed by any AI vendor. Product names are used only to say which tool a lesson teaches.
How is this page different from Nightschool for Teams?
This page helps one person choose AI training for their own work. Nightschool for Teams is for managers who want to set up the same training for a group, with invite codes and onboarding.
What is the difference between AI literacy and AI training?
AI literacy is knowing what AI assistants are, how they produce answers and where they fail, such as stating a wrong fact in a confident tone. AI training builds on that with practice: you do real tasks with an assistant and learn to check the result. Good training covers both. Our first Claude course has a lesson on what the assistant does well and where it falls short.
Does everyone need the same AI training?
The foundation is the same for everyone: giving context, picking suitable tasks, working in rounds and checking the output. After that it should follow the job. A finance analyst, a recruiter and a nurse hand different tasks to an assistant, so practice on your own work, and use our guides by role and industry to see what people in your field use it for.
Is AI training only for people in tech jobs?
No. Most of what people use AI assistants for at work is writing, reading and organizing: emails, reports, summaries, meeting notes and research. Our guides cover roles such as sales, HR and operations, and industries such as banking, insurance, healthcare and pharma.
How do I know if AI training is working for me?
Pick two or three tasks you do every week. Before you start, note how long each one takes and how often you have to fix something. After a few weeks of using the assistant on them, compare. If the drafts need fewer corrections and you can explain how you checked each one, it is working.
What languages are the courses in?
This page and the starter courses for Claude, Microsoft 365 Copilot, ChatGPT and Gemini are available in English and 13 other languages, including Spanish, German, French, Japanese, Korean, Hindi and Portuguese. Use the language menu at the top of the page to switch.
How do you keep AI training current when the tools change?
AI tools change every few months, so a course that was right last year can be out of date now. Our What's New in Claude course adds lessons as Anthropic ships new Claude features. Whichever course you pick, check that it was updated after the tool's most recent major change.
Keep reading
A self-paced starting point for people with a day job.
Every course, in order, from first lesson to agents.
How to fit an AI course around a full-time job.
Realistic time ranges for practical use.
Prepare for official certifications offered by AI vendors.
How our lessons are made, checked and kept current.
Start with one real task
The free first Claude lesson takes about 6 minutes, and you don't need a card.
Nightschool AI is independent. It is not affiliated with or endorsed by Anthropic, Microsoft, OpenAI, Google or Meta. Product names on this page only identify the tools a course teaches.