How to Use an AI Assistant for the First Time (2026)

You are reading Chapter 2 of 36 in the AI Bootcamp, AI Central's free course for professionals who want to use AI in their own work. The chapters build on each other, so if this is your first one, Chapter 1 is the place to start. All of them sit on the AI Bootcamp page, and three more arrive every week.

Using an AI assistant for the first time takes about ten minutes: open one of the four main assistants, sign in, and type a real task from your week into the box. The hard part is not the account. It is knowing what you are looking at once you are through the door, and what to say first.

This guide shows you the three parts of the screen that matter on day one, walks through account creation on ChatGPT, Claude, Gemini and Copilot, and gives you the one message worth sending first.

We publish AI systems like this for 300,000+ senior professionals at AI Central, and more of them live in the AI Central Library.

Most people open an assistant, type "what can you do", read a paragraph about writing and analysis and brainstorming, and close the tab. Nothing was wrong with the tool. The question was a test, and a test question gets you a list of things it can do rather than a piece of your own work done.

There is a second reason first sessions stall. The screen is busier than it needs to be. Model pickers, buttons, a sidebar, a menu of things you have no reason to touch yet. It looks like something you are supposed to configure before you start, so people assume there is a setup step they have missed. There isn't one.

If you have not picked your assistant yet, Chapter 1 picks one for you in five lines. Come back here once you have.

You don't need four different guides for four different tools. Every one of these accounts opens the same way, and the only real difference is what each one asks you for.

Go to chatgpt.com. Sign up with email, Google, Microsoft or Apple. No phone number is required for a normal signup.

Go to claude.ai. Sign in with email or Google, then Claude asks for a real mobile number and texts you a code. This step can't be skipped, and it won't accept a Google Voice number, a VoIP number, or a landline: it needs an actual mobile phone.

Go to gemini.google.com. Sign in with the same Google account you already use for Gmail. That's the entire process.

Go to copilot.microsoft.com. It's free, no subscription needed, but sign in with a Microsoft, Apple or Google account.

If you want the longer version of this step, with plans, pricing and privacy settings side by side, read our setup guide for ChatGPT, Claude and Copilot.

Once you're through the door, the sequence is identical everywhere.

You'll see a sign-in screen.

You land in a blank chat.

You get a real, custom reply.

Steps one and two are about getting in the door. Step three is the only one that actually matters.

Every one of these four screens is busier than it needs to be for a first-timer. Three things are worth finding. The rest can wait.

The box you type into. On day one it is the entire product.

Every past chat is saved and searchable. Start a new chat when you change topics, so one thread stays one topic and your history stays useful.

This is the one that differs by tool. ChatGPT gives you one model plus a Think button. Claude has one fixed model and no dial at all. Gemini has a real three-way Fast, Thinking and Pro picker. Copilot runs on Auto by default, with a deeper option if you ask for it.

None of this is worth optimizing on day one. Defaults exist because they work for most questions most of the time.

The fastest way to actually understand what this tool is for is to stop testing it and start using it. Three openings that work:

Explain something you already half-know. Write a first version of something you were about to write anyway. Check something you already wrote.

Take the task you circled in Chapter 1 and type it in, in your own words. You do not need special wording, a framework, or a magic phrase. Whatever comes back will not be perfect, and that is fine. Your only job in this chapter is to have one real conversation and see what it is actually like.

Before: you have an account you opened once and a screen you are slightly wary of, so the tab stays closed and you keep meaning to come back to it.

After: you are signed in, you can name the three parts of the screen that matter, and you have had one real conversation about your own work. You have used it, not just signed up for it.

Two things carry into Chapter 3: your signed-in account, and one more real task from your work this week. If you want a one-page reference to keep beside you while you work, start with our free AI cheat sheets, and the rest of the syllabus lives in the AI Central Library.

No. All four have a free tier that is enough for this course. Copilot is free with a Microsoft, Apple or Google sign-in, and Gemini works with the Google account you already have.

For Claude, yes. It asks for a real mobile number and texts a code, and it will not accept a Google Voice number, a VoIP number, or a landline. ChatGPT does not require a phone number for a normal signup. Gemini uses your existing Google account, and Copilot uses a Microsoft, Apple or Google account.

The default. ChatGPT has one model plus a Think button, Claude has one fixed model, Gemini has a Fast, Thinking and Pro picker, and Copilot runs on Auto. Defaults exist because they work for most questions most of the time, and none of this is worth optimizing before you have had a first conversation.

A real task from your own week rather than a test question. Explain something you already half-know, write a first version of something you were going to write anyway, or check something you already wrote. Plain words are enough. You do not need a framework or a special phrase.

Start a new chat when you change topics. Every past chat is saved and searchable, so one thread per topic keeps your history worth searching later.