Master ChatGPT for Work is a free ChatGPT course for people who already use ChatGPT at the office and know they are only getting a fraction of what it can do. It has 12 short modules in a set order, from keeping client data out of training to handing ChatGPT a multi-step job. Each module is a real office task, and you practice every click in a copy of ChatGPT in your browser before you try it in your own account. You can start the course here or follow it as a learning path.
This page is for office managers, owners, team leads and staff at accounting firms, agencies, law offices and other white-collar teams. It explains what the course covers, why the order matters and which modules need a paid plan in your real account. The examples use made-up businesses and fake client data only.
Why learn ChatGPT this way?
Most teams learn ChatGPT one tip at a time. Someone shares a prompt in Slack, someone else finds a setting, and nobody knows which habits are safe or which features are worth paying for. The result is the same everywhere: people paste client data without checking the settings, re-explain their job in every chat, lose the thread halfway through a long task and send answers nobody fact-checked.
The cost is easy to see. If five people each lose 20 minutes a day to retyping context, fixing vague answers and redoing work, that is more than eight hours a week. One made-up figure in a client email can cost more than that. The fix is not a cleverer prompt. It is a short list of skills, learned in the right order, that turns ChatGPT from a chat window into a tool that knows your job, reads your files and does multi-step work while a person still signs off.
That is what this course does. Each module starts with the business problem and the payoff, then shows the feature that solves it.
Who is this ChatGPT course for?
- Office staff who write emails, recap meetings and work in spreadsheets every day.
- Managers and team leads who want their team to use ChatGPT the same safe way.
- Owners deciding whether a paid ChatGPT plan is worth it, and for what.
You do not need any coding or technical background. If you have never used ChatGPT at all, start with AI at Work: Start Here first, then come back.
The 12 modules, in order
The course goes beginner to advanced. The first four modules make every chat safer and better. The middle four put ChatGPT to work on files and real tasks. The last four use the newest features, most of which need a paid plan in your real account.
| # | Module | The business problem it solves | Plan needed in your real account |
|---|---|---|---|
| 1 | Keep client data out of training in 5 minutes | Client details end up in a tool that may learn from them | Any |
| 2 | Stop re-explaining your job: memory vs custom instructions | Every new chat starts from zero | Any |
| 3 | Long jobs without lost context: new chat or same chat | Long chats drift and forget the brief | Any |
| 4 | Prompts that work the first time | Vague answers that need three rewrites | Any |
| 5 | Turn a meeting into minutes and tasks | Nobody writes up the meeting, so tasks get lost | Any |
| 6 | Catch made-up facts before a client sees them | Confident answers that are wrong | Any |
| 7 | Messy export to answers and a chart | A spreadsheet nobody has time to read | Any |
| 8 | One home per client: Projects | Files and instructions scattered across chats | Any |
| 9 | Save your custom GPT before it retires (Enterprise: December 11): move it to a plugin | A GPT your team relies on is being retired | You must be the GPT's creator or a workspace admin |
| 10 | Vendor or competitor brief with Deep Research | Hours of research before a buying decision | Higher limits on paid plans |
| 11 | Hand off a multi-step task with ChatGPT Work | Copy-paste jobs across several sites and files | Plus or higher on the web (desktop app: any plan) |
| 12 | A tracker that updates itself: Space, Pages and Dots | Recurring checks that someone forgets | Pro or Business Premium |
Modules 2, 9 and 10, and the Space part of module 12, are being built now and appear in the course as soon as each one is live. Everything that is linked above is free and ready today.
Modules 1 to 4: make every chat safe and useful
Start with ChatGPT data controls, because everything else in the course involves real work. You will turn off model training for your account, use a temporary chat for sensitive one-off questions and learn what still gets stored. Then module 2 covers the difference between ChatGPT memory and custom instructions: what each one remembers, where to manage it and how to set it up for a made-up firm so ChatGPT stops asking who you are.
Module 3 answers a question people search for every day: should you start a new chat or keep going? You will learn when a long chat starts to lose the brief and how to write a handoff summary that carries the work into a fresh chat. Module 4 is the prompt habit that fixes most bad answers: say the role, the context, the task and the format, and put pasted text between clear markers.
Modules 5 to 8: put ChatGPT to work on real files
Module 5 turns a meeting into minutes, decisions and an action table with owners and dates. Module 6 is the safety net: how to ask ChatGPT to show its sources, when to turn on search, and a short check that catches a hallucination before it reaches a client. Module 7 takes a messy sales export with a bad header row and turns it into answers and a chart you can download. Module 8 gives each client one Project with its own files and instructions, so every chat about that client starts with the right context.
Modules 9 to 12: the newest features, with a person in charge
These modules use features that change fast, so each one is dated and checked against OpenAI's own documentation when it is built.
Save your custom GPT before it retires (Enterprise: December 11)
OpenAI is retiring custom GPTs and replacing them with plugins. A plugin is a package for a workflow that can include skills (written instructions for how to do a task) and apps (connections to other services). OpenAI's custom GPT retirement FAQ lists December 11, 2026 as the scheduled retirement date for Enterprise workspaces and says other plans may follow the same timeline. After that date, OpenAI's migration guide says custom GPTs "stop running and leave the GPT directory."
Here is what OpenAI's migration guide says to do, checked on October 17, 2026:
- Go to My GPTs, find your GPT under Created by me, and select Migrate to plugin. The option shows only to the GPT's creator or a workspace admin, for a published GPT, when plugins are turned on.
- Migration turns the GPT's instructions into a skill and brings over its knowledge files and connected apps.
- Rebuild anything that does not carry over: custom actions do not transfer, conversation starters do not transfer one-to-one, and the selected model does not carry over.
- Test the plugin with a few prompts you know the GPT handled well, including one harder case, and compare the results.
- After migration the original GPT becomes read-only, and it keeps working until retirement, so make your last edits before you migrate.
Module 9 walks through those steps on a made-up company's GPT, then re-tests it as a plugin. If you would rather not wait, you can also move a custom GPT's job into a ChatGPT Project today.
The rest of the advanced modules
Module 10 uses Deep Research to write a cited brief on three made-up software vendors, then exports it. Module 11 hands ChatGPT Work a multi-step job with connected apps and an approval step before anything is sent. Module 12 sets up a tracker that updates itself with Space and Pages, then gives a ChatGPT dot a recurring check, with read-only access and a person approving every action.
How each module works
Every module follows the same pattern, so you always know what comes next:
- Why it matters. The business problem and what fixing it is worth.
- Practice. Every step in a copy of ChatGPT in your browser. It checks your work, and nothing touches your real account.
- Do it for real. A short checklist to repeat the task in your own ChatGPT, with sample files where the module needs them.
- Quick check. Three questions to make sure it stuck.
Most modules take 5 to 15 minutes. You can stop anywhere and pick up where you left off.
What you need
- A ChatGPT account. The practice copies need no account at all, and modules 1 to 8 work on the Free plan in your real account (Projects and file uploads have lower limits on Free).
- For modules 9 to 12, the plan each module names: ChatGPT Work on the web needs Plus or higher, Dots need Pro or Business Premium, and migrating a GPT needs its creator or a workspace admin.
- No real client data. Use the sample files and made-up businesses in each module until you are sure of your settings.
Key takeaways
- Learn ChatGPT in order: data controls and memory first, then prompts, files and Projects, then plugins, research and agents.
- Every module is a real office task with a clear payoff, practiced in a safe copy of ChatGPT.
- Custom GPTs are being retired. OpenAI's FAQ gives December 11, 2026 as the Enterprise date, so move the ones you rely on to plugins before then.
- The newest features need a paid plan in your real account, and each module says which.
Frequently asked questions
- Is this ChatGPT course for work free?
Yes. Every module is free, and the practice runs in your browser without a ChatGPT account. You only need a paid ChatGPT plan to repeat the advanced modules in your own account.
- How long does it take to master ChatGPT for work?
Each practice takes about 5 to 15 minutes. The point is not speed: do one module, use it on real work for a day, then do the next.
- Do I need a paid ChatGPT plan?
Not for modules 1 to 8, which work on the Free plan with lower limits. The advanced modules use features that are limited or missing on Free in your real account, such as ChatGPT Work on the web (Plus or higher) and Dots (Pro or Business Premium). You can still do every practice copy for free.
- What happens to my custom GPTs?
OpenAI is replacing custom GPTs with plugins. Its FAQ lists December 11, 2026 as the scheduled retirement date for Enterprise workspaces and says other plans may follow. The creator of a published GPT, or a workspace admin, can use Migrate to plugin from My GPTs. Custom actions have to be rebuilt.
- Is it safe to use ChatGPT with client data?
Only after you check your settings, which is why module 1 is data controls. Use the free practice copies and fake data first, follow your firm's AI policy, and never paste anything your clients would not want shared.
Sources
- OpenAI: Moving your custom GPT workflows to plugins (checked October 17, 2026)
- OpenAI Help Center: Custom GPT retirement and migration FAQ
- PCWorld: Custom GPTs in ChatGPT are going away. Here's how to save yours (September 18, 2026)