Author: Pete

A contract lands in your inbox at 4 pm. It is nine pages, full of “hereinafter”, and the reply is due tomorrow. You are not a lawyer, a lawyer costs money you would rather not spend on a routine document, and a small voice says maybe an AI can help.

Here is the short answer. You can use AI for legal documents to understand them, summarize them, and prepare your questions. You should not use it as your lawyer or to decide whether to sign. Let the AI read first, then you read the parts that matter, and bring anything with real money or risk attached to a qualified professional.

Want a free starter path for using AI like this in your daily work? Start on the RobotsHelp home page.

What AI is good at with legal documents

Legal text is long, formal and repetitive. That is exactly where a language model saves you time. Four jobs work well:

• Plain-English summary. What does this document ask of me, and what do I get?
• Finding the clauses that matter. Payment, termination, renewal, liability, confidentiality, who owns the work.
• Explaining one confusing sentence. Paste the clause and ask what it means in everyday words.
• Making a list of questions. You walk into the conversation with the other side, or with a lawyer, already knowing what to ask.

If you have not used an assistant on long files before, our guides on how to summarize long documents with ChatGPT and how to analyze long PDF files with Claude cover the basics of getting a file in.

What AI is bad at, and why it matters here

An assistant can sound sure of itself and still be wrong. In legal work that is the dangerous part, because a wrong answer reads exactly like a right one. Keep these limits in mind:

• It does not know your situation. Rules differ by country and by state, and a clause that is normal in one place can be unusual in another.
• It can invent things. Never trust a case name, a law number or a quotation it gives you until you have opened the real source yourself.
• It can miss what is not on the page. A good lawyer notices the clause that should be there and is not. A summary tool only describes what it was given.
• It cannot take responsibility. If you sign, the decision is yours.

Professionals are told the same thing. The American Bar Association’s first ethics guidance on generative AI says lawyers must understand the benefits and risks of the tools they use and keep client information confidential. Read the summary here: ABA issues first ethics guidance on a lawyer’s use of AI tools. If the people whose job it is to read contracts are told to check the output, so should you.

Before you paste anything: privacy first

A contract often contains names, prices and details the other side expects you to keep private. Before you upload it, do two things.

1. Check your settings. On personal ChatGPT plans, OpenAI’s help page says sharing your chats to improve its models is on by default, and you can switch it off under Settings, then Data controls. The same page says business and enterprise workspaces are not used for training by default. Details: Data controls in ChatGPT (OpenAI Help Center). Anthropic explains its own rules for Claude here: Is my data used for model training? (Claude Help).
2. Remove what the AI does not need. Swap names for “Company A” and “Company B”, and blank out bank details, ID numbers and addresses. The summary will be just as useful.

If your contract says it is confidential, or your employer has an AI policy, follow that first. When in doubt, ask.

A safe workflow in five steps

This is the order I would follow, and it keeps you in charge of every decision.

1. Clean the document. Remove or replace names and private numbers. Use the text, not a blurry scan.
2. Ask for a plain-English summary first. You want the big picture before the details.
3. Ask for the clauses that matter. Have it list each one and quote the exact words, so you can check it against the original.
4. Verify every quote. Search the document for the words the AI quoted. If you cannot find them, the AI made something up. Throw that part away.
5. Write your questions, then decide who answers them. Some you can ask the other side. Some need a professional.

Prompts you can copy

Paste the cleaned text under any of these. Change the words in brackets.

1. The plain-English summary

2. The clause finder

3. The “what should worry me” check

4. The two-version comparison

The comparison prompt is the one I would keep. When the other side sends back “just a few small edits”, it is a fast way to see whether they really are small. Still open the changed clauses and read them yourself.

What to bring to a lawyer instead

Some documents are not a DIY job, even with a good summary in hand. Bring in a professional when:

• A lot of money is involved, or the contract runs for years.
• It touches employment, property, company ownership, or anything that could end up in court.
• The other side has their own lawyer.
• The summary left you more worried than before.

Here the AI still earns its keep. A lawyer who gets a clean summary and a list of your questions can work faster, and that usually means you pay for less of their time. Ask them how they prefer to receive it.

Which assistant should you use?

For reading and explaining, the main assistants all do a reasonable job on a cleaned, text-based document. Pick the one you already use and know how to steer. Your prompt and your checking matter more than the brand. If you want help writing sharper prompts, see our best Claude prompts for work.

If you sign many contracts and think about a dedicated tool, read how to pick AI contract review software, or skip it first.

Common questions

Can AI replace a lawyer for simple documents?

No. It can help you understand a simple document and prepare, but it does not know your local rules and it cannot be held responsible for being wrong. Treat it as a very fast reading assistant.

Can I ask AI to write a contract from scratch?

You can ask for a first draft to react to, and it is useful for seeing what a document usually covers. Do not send or sign it as is. Have someone qualified review anything that creates real obligations.

Is it safe to upload a confidential document?

It depends on your plan, your settings and your agreement with the other side. Remove names and private numbers, check the privacy settings linked above, and follow your employer’s rules.

What if the AI and the document disagree?

The document wins. Always. Go back to the original words and read that clause yourself.

Where to go from here

Use AI to read faster and ask better questions, and keep the judgment for yourself. Next time a long document lands in your inbox, you can have a summary and a list of questions ready in minutes, and walk into the conversation calm instead of rushed.

Want more practical ways to put AI to work for you? Join the free Launchpad below.

This article is general information, not legal advice.

You saw a video promising a thousand dollars a day with ChatGPT, and part of you wants it to be true. Here is the honest version. You can use ChatGPT to make money, but it works as a tool that makes your existing skills faster, not as a machine that earns while you sleep.

The short answer: pick a service people already pay for (writing, editing, customer replies, bookkeeping notes, translation checks, research), use ChatGPT to produce the first draft in minutes, then add your own judgment and finish the work properly. The money comes from the client and the quality, not from the chatbot. Below are six ways that are real, what each one needs, and the traps to avoid. Want a free place to start learning practical AI at work? Join through the RobotsHelp home page.

What “making money with ChatGPT” really means

ChatGPT does not have customers. You do, or you will. It helps in three places: finding the work, doing the first 70 percent of it faster, and handling the admin around it. If you already do something a client values, you can do more of it in the same week. If you have no skill clients value yet, ChatGPT can help you learn one, but it will not replace the learning.

Be careful with anyone selling a shortcut. The US Federal Trade Commission has sued business-opportunity sellers who used “AI” to promise big, easy online income, and its consumer advice says to ask for the legal disclosure documents and treat conflicting promises as a warning sign (FTC consumer alert on Operation AI Comply). A good rule: if the pitch starts with the income and not with the work, walk away.

Six honest ways to use ChatGPT to make money

1. Freelance writing and editing, done faster

Clients pay for clear, correct, on-brand text. ChatGPT is good at a rough first draft and at tightening a messy one. You supply the facts, the voice and the final read. Start with one narrow type of text, such as product descriptions for small shops or newsletters for local businesses, and learn it well. Our guide to AI for content creation shows how to turn one idea into a week of posts.

2. Social media management for small businesses

Many owners know they should post and never do. You can offer a monthly package: plan the topics, draft the posts with ChatGPT, adjust them to the owner’s voice, and schedule them. The prompts in ChatGPT prompts for a social media manager are a good starting kit.

3. SEO briefs and content outlines

Writers and agencies need briefs: the search question, the headings, the points to cover. ChatGPT can draft the structure while you check it against what really ranks. See ChatGPT prompts for an SEO content brief for a ready method.

4. Customer support replies and help-desk setup

Small companies lose hours answering the same ten questions. You can write their answer library with ChatGPT, test it on real past emails, and hand over a tidy document. Start with ChatGPT prompts for a customer support agent, and if the client wants more, look at how to automate customer support with AI agents.

5. Research, summaries and reports

Busy people pay for “read this and tell me what matters.” ChatGPT can summarize long documents and pull out key points. You verify every number against the source. This is also where mistakes hurt most, so never send a figure you have not checked yourself.

6. Be the AI helper inside your current job

The easiest money is often your own salary. Use ChatGPT to finish reports, emails and prep work faster, then take on the work that grows your role. This needs no new clients and no new platform, and it is the best first test of whether a task suits AI. Our overview of how to use Claude for work and how to use Gemini for work applies here as well.

A simple first-week plan

Do not start with a business plan. Start with one test.

1. Day 1: List three things you can already do that someone would pay for. Pick the one you enjoy most.
2. Day 2: Write a one-line offer: who it is for, what they get, how long it takes.
3. Day 3: Build a sample. Use the prompt below, then edit the result until you would be proud to show it.
4. Day 4 and 5: Send the sample to five people who could use it, or offer it to one past colleague.
5. Day 6 and 7: Note what they asked about. Adjust the offer. Only then think about price.

A prompt to turn your skill into a sample:

Does it work with Claude and other tools?

Yes. The same ideas work if you prefer Claude or Gemini: the tool changes, the method does not. Choose one assistant, learn its strengths, and keep your prompts in a document so you can reuse them. If you are choosing between them, see ChatGPT vs Claude.

The rules that keep you out of trouble

You can use the output commercially, with limits. OpenAI’s terms say that as between you and OpenAI, you own the output, and OpenAI assigns you its rights in it. The same terms say output may not be unique, that you must check it for accuracy, and that you may not claim output was human-generated when it was not (OpenAI Terms of Use). Read them once, they are short.

Check what your client and platform allow. On Upwork, for example, freelancers may use generative AI tools unless the client expressly restricts it, and the platform recommends talking about tools and expectations with the client up front (Upwork’s April 2023 AI update). Other platforms and contracts differ, so read yours.

Protect your client’s private information. Do not paste confidential contracts, customer lists or passwords into any chatbot unless your client has agreed and you understand the tool’s data settings.

Never promise results you cannot control. Say what you will deliver and by when. Do not promise rankings, sales or followers.

Common mistakes

• Selling raw ChatGPT output. Clients can get that for free. They pay for your checking, your taste and your reliability.
• Chasing ten services at once. One narrow offer beats a menu.
• Skipping the fact check. ChatGPT can sound sure and be wrong. One bad number costs you a client.
• Buying a course before testing the idea. Make one sample first. It costs nothing.

Questions people ask

How much can you earn with ChatGPT?

There is no honest single number. It depends on your skill, your clients and your hours. Anyone quoting a daily income without showing the work behind it is selling something.

Do I need to pay for ChatGPT?

Not to start. A free account is enough to build your first sample and test an offer. Upgrade only when limits get in your way.

Is it legal to sell work made with ChatGPT?

Generally the terms allow commercial use, as covered above, but you stay responsible for accuracy and for not infringing anyone’s rights. Check your own local rules and your client contract.

What is the fastest way to start?

Offer a task you already know how to do, speed it up with ChatGPT, and show one sample to a real person this week.

Your next step

Pick one skill, one type of client and one sample. That is the whole plan for this week. If you want short, practical routines for using AI in real work, join the free Launchpad below.

Thinking about credentials first? See what the ChatGPT certification course really offers.

You have a photo that is almost right. The background is messy, there is a stranger walking through the frame, or the colours look flat. You do not want to learn a design program for one picture.

Here is the short answer. Open Gemini, add your photo with the plus button, then type what should change and what must stay the same. Check the result, then ask for small fixes in the same chat. That is the whole method to use Gemini to edit photos, and the rest of this page shows how to do it well, with prompts you can copy.

Want a free starter path for using AI like this in your daily work? Start on the RobotsHelp home page.

What you need before you start

Three things, and they are easy to miss.

• A Google account signed in to the Gemini app. It works on the Gemini website and in the mobile app.
• Age 18 or over. Google’s help page says editing images is not available to users under 18, and that school accounts can have different restrictions.
• A supported country and language. Google notes that availability depends on both. If you do not see the image tools, that is the first thing to check.

Source: Generate & edit images with Gemini Apps (Google help). Google changes names and limits from time to time, so if a button looks different on your screen, trust your screen over any article, this one included.

How to edit a photo in Gemini, step by step

This is the basic flow. It takes a couple of minutes the first time.

1. Open Gemini in your browser or the mobile app and start a new chat.
2. Add your photo. Tap the plus or Add files button, then pick Photos, Camera or Files. Google’s help page says each file can be up to 100 MB and you can add several files to one prompt.
3. Type what to change. One clear request is better than five at once.
4. Submit and look closely. Zoom in on faces, hands, text and edges. That is where mistakes hide.
5. Ask for a fix in the same chat. Say “keep everything, but make the sky a little less bright” and Gemini works from the picture it just made.
6. Save the one you like. Download it or save it to your device before you close the chat.

You can also pick the image tool from the tools menu (it is labelled Create images) and upload a photo there. Both routes lead to the same kind of edit. If you have already generated an image in Gemini, you can open it from your Library and tap Chat under the image to keep editing it.

If you are new to the app itself, my Gemini tutorial for beginners walks through the basics first.

The prompt formula that keeps your photo intact

Google’s own prompting guide for its image model says editing needs a different mindset than creating. You already have a base image, so your prompt should say what is changing and what is staying the same. Source: Google Cloud’s prompting guide for Nano Banana.

I use a simple three-part sentence:

[Verb] + [the one thing to change]. Keep [what must stay the same].

Start with a strong verb: replace, remove, add, change, brighten. Describe what you want, not what you do not want. Google’s guide gives “empty street” as a better wording than “no cars”. And always say what to keep. Without that line, Gemini sometimes changes more than you asked for.

Copy and paste prompts for common photo jobs

Change the details to match your picture. Each one asks for a single job.

Replace a messy background

Replace the background with a plain white studio backdrop. Keep the person, the pose and the lighting on them exactly the same.

This is close to the example Google’s help material gives for a background swap. Good for a headshot or a product picture for a listing.

Remove a person or object

Remove the man in the red jacket on the right. Fill the space so it looks natural and keep everything else unchanged.

Be specific about which person or object. “Remove the stranger” is vague if there are three strangers.

Fix dull light and colour

Brighten the photo slightly and add a warmer, natural colour tone. Keep the faces, the composition and all details the same.

Words like “slightly” help. Without them the edit can go too far.

Change the style

Turn this photo into a watercolor painting. Keep the same people, the same composition and the same colours as the main palette.

Google’s guide describes style transfer like this: upload a photo and ask for the same content in a different artistic style.

Combine two images

Use the first photo as the base. Add the plant from the second photo on the desk, on the left side. Match the lighting.

Upload both pictures in one prompt. Google’s help page says you can upload several images and ask Gemini to create a new image from them.

Make a cleaner photo for work

Make this product photo look clean and professional: neutral light grey background, soft shadow under the product. Keep the product shape, colour and label exactly the same.

For anything with text on it, check the lettering carefully afterwards. If a label has changed, ask again and say “keep the label text exactly as in the original”.

Edit in small steps, not one big prompt

This is the habit that made the biggest difference for me. A prompt with six changes gives you a picture that is a little wrong in six places. Three short prompts in a row usually give you a picture that is right.

Try this order: first the big thing (background or removal), then light and colour, then the small details. After each step, look at the result. If a step went wrong, say what was wrong in plain words and ask Gemini to try again from there. Google’s guide calls this iterating: refine the image with follow-up prompts, like a conversation.

Where Gemini photo editing goes wrong

It is good, but it is not magic. Know these limits so they do not surprise you.

• Faces can shift. If the edit is not about the face, say “do not change the face” and check it anyway.
• Text in the image can change. Signs, labels and screens are easy to damage.
• Hands and small details can look off. Zoom in before you use the picture anywhere.
• Quality depends on the model. Google’s help page describes a faster Lite option that is not designed for multiple image references or multiple edits. If you need that, switch to the Flash or Pro model option.

None of this means you should avoid it. It means you are the editor. Gemini does the first 80 percent quickly, and you do the final check.

Is it okay to edit other people’s photos?

Use your own photos, or photos you have the right to change. Do not edit a photo of a real person in a way that could mislead others about what they said, did or looked like. For work, that also means being honest when an image has been changed. If a client or colleague would feel tricked, do not do it.

Connect Google Photos (US only, for now)

Google’s help page describes a Google Photos connection. You turn on Google Photos under Connected apps in Gemini’s settings, and then Gemini can reach your photos without you uploading each one. Google says this is only for users 18 and over in the US, signed in to a personal Google account, and it is not available in some US regions. If you are not in the US, skip this and upload the photo directly.

Photos are one design job; a logo is another. For that, read our guide to the best AI tool for logo design.

How to use this for real work

Photo editing is a small job, but it saves real time in a normal week. A few places I would start:

• Profile and team photos: a cleaner background for a headshot you already have.
• Product and listing photos: neutral background, better light, no reshoot.
• Slides and documents: change the style of a photo so it fits your deck.
• Social posts: remove distractions from the frame before you post.

If you mostly want pictures made from scratch, not edited, the comparison in best AI tool for image generation will help you pick. And if you want Gemini to help with the rest of your day, see how to use Gemini for work and how to use Gemini in Gmail.

Quick answers

Can Gemini edit a photo I upload?

Yes. Google’s help page lists uploading an image and asking Gemini to make edits as one of the ways to edit images.

Do I need a paid plan?

Plans and limits change, so check the current Gemini and Google AI plan pages for what your account includes. I will not guess at limits here.

Why did Gemini change parts of my photo I did not mention?

Usually the prompt did not say what to keep. Add a line like “keep everything else exactly the same” and try again.

Can it add or remove things?

Yes. Google’s guide describes adding elements from a second image and removing items by describing them in the prompt.

Your next step

Pick one photo you already have. Write one prompt with a change and a keep line. Look at the result, then fix one thing. That is ten minutes, and you will know whether this fits your work.

When you want more small, practical wins like this, join the free Launchpad below.

Starting from scratch instead of an existing photo? Here is how to use Gemini to create images for work.

You know you should post more. The ideas are there, in your head, in client calls, in the questions people keep asking you. But every time you sit down to write, the page stays empty and the week fills up with real work. So nothing goes out, and you feel a bit more behind than last week.

Here is the short answer. The best way to use AI for content creation is not to ask it for “a post about X”. It is to feed it your raw material (a voice memo, call notes, a question a client asked), let it shape that into one solid main piece, and then let it cut that piece into smaller ones for LinkedIn, your newsletter and a short video. You stay the source and the final editor. AI does the blank page, the structure and the reformatting. Below is the full workflow, with the prompts, step by step.

Want a free starter path for using AI like this in your daily work? Start on the RobotsHelp home page.

What AI is good at in content creation, and what it is not

Be honest about the split before you start. It saves you from the two classic mistakes: expecting AI to have your ideas, or not trusting it with anything at all.

AI is good at:

• Turning messy notes or a rambling voice memo into a clear outline.
• Giving you a first draft to react to, which is much easier than starting from nothing.
• Rewriting one piece into other formats: a post, an email, a 60 second script.
• Offering ten headline options so you can pick the one that sounds like you.
• Spotting gaps: “what would a beginner still not understand here?”

AI is not good at:

• Knowing what happened in your work this week. Only you have that.
• Having a real opinion. It averages. Your point of view is the part people follow you for.
• Being right about facts every time. It can state wrong numbers with total confidence, so you check every claim.
• Sounding like you without help. Out of the box it sounds like everyone else.

So the workflow below is built around one idea: you bring the substance, AI brings the speed.

Step 1: Give the AI your voice once, not every time

Most people paste “write in a friendly tone” into every chat and get the same flat result. Do this once instead. In Claude you can create a project with its own instructions and files; Anthropic describes projects as self-contained workspaces with their own chat histories and knowledge bases, where you can upload documents and set project instructions (Claude Help Center: What are projects?). ChatGPT has a similar setup with projects and custom instructions. Either works.

Put three things in it:

1. Three to five pieces you wrote yourself and liked. Old posts, a good email to a client, anything that sounds like you on a good day.
2. A short voice file: who you write for, words you use, words you never use.
3. Your topics: the four or five things you want to be known for.

Here is a voice file you can copy and fill in:

That last line matters. It turns the AI from a guesser into a helper that asks questions.

Step 2: Start from your raw material, not from a topic

“Write a post about cash flow tips” gets you the same post the internet already has a thousand times. Raw material from your own work gets you something only you could publish.

Good raw material is easy to collect during a normal week:

• A question a client asked you on a call.
• A mistake you saw someone make (no names).
• A five minute voice memo on your walk: “Here is what I wish people knew about X.”
• A before and after: how you used to do a task, how you do it now.

Keep a simple note on your phone called “content ideas” and drop these in as they happen. Once a week, pick one. If you record a voice memo, most phones can transcribe it, or you can paste the audio transcript straight into the chat.

Then use this prompt:

“Do not write the article yet” is the trick. You get to fix the thinking before any polished text hides the weak spots.

Step 3: Write one main piece, properly

Pick one format as your main piece. For most professionals that is a blog article or a newsletter issue, because it is the longest and everything else can come out of it. If you are more comfortable on camera, it can be a short video script instead.

Once the outline looks right, ask for the draft:

The [CHECK] instruction keeps invented facts out of your draft. Fill those spots yourself, with a real source, or cut them.

For the deeper version of this step, we have separate guides for each format: writing a blog post with AI that sounds like you and writing a newsletter with AI.

Step 4: Turn one piece into five

This is where AI saves the most time. Repurposing by hand is boring work that most people skip. With AI it takes minutes, and it means one good idea shows up in several places where your audience already is.

Paste your finished main piece and use:

“Using only what is in it” is important. It stops the AI from adding new facts in the smaller pieces, which is where errors often slip in, because you check them less carefully.

Each format has its own habits. A LinkedIn post that works looks very different from a good email. Our guide on using AI to write LinkedIn posts that sound like you goes into that platform in detail, and if you plan short videos, AI UGC videos for small business covers the video side.

Step 5: The edit pass that makes it yours

Never publish the first draft. This is the step that separates content people trust from content they scroll past. It takes ten to fifteen minutes and you can do it with a coffee.

Run through this list for every piece:

1. Read it out loud. Anything you would never say to a client, rewrite or cut.
2. Add one real detail. A number from your own work, a short story, a client situation (anonymous). This is the part AI cannot give you.
3. Check every fact and number. If you cannot find a source in two minutes, take it out.
4. Delete the filler. Lines like “Let’s dive in”, “In this post we will explore”, and summaries that repeat what you just said.
5. Sharpen your opinion. Where the draft says “it depends”, say what you actually recommend.

You can also ask the AI to help with the edit, which works surprisingly well:

Asking for a list instead of a rewrite keeps you in charge of the final words.

How to use AI for content creation without sounding like a robot

This is the worry most people have, and it is fair. Readers can feel generic writing quickly, even if they cannot say why. The fixes are simple:

• Start from your raw material (Step 2). Generic input gives generic output.
• Use your voice file and real samples (Step 1). Telling the AI “be casual” does far less than showing it how you write.
• Ban the usual phrases. Put the words you hate in the voice file so they stop showing up.
• Vary your sentences in the edit. AI drafts often have the same rhythm line after line. Break a long sentence in two. Add a short one. Like this.
• Keep the personal parts human. The story, the opinion, the “here is what I got wrong” line: write those yourself.

Which tool to use matters less than this process. If you want to compare, our side by side of ChatGPT vs Claude for writing shows how the two handle tone and longer drafts.

A simple weekly rhythm

A workflow only helps if it fits into a real week. Here is a rhythm that works for a busy professional who wants to post consistently without it eating their days. Adjust the days to your own calendar.

That is under two hours for one main piece plus several smaller ones. The comments and replies are the part you never hand off. They are where the real relationships start, and very often where next week’s idea comes from.

If you manage social channels as part of your job, our weekly prompt kit for social media managers fits right into Wednesday.

The rules you should know

Two short points so you do not get caught out.

Search engines. Google has said plainly that “appropriate use of AI or automation is not against our guidelines”, as long as it is not used to produce content mainly to manipulate search rankings (Google Search Central: guidance about AI-generated content). In plain words: helpful content made with AI is fine. Pumping out dozens of thin articles to game Google is not. The workflow above, one real idea and a careful edit, is the first kind.

Video platforms. YouTube requires creators to disclose AI-generated or altered content that seems realistic, for example making a real person appear to say something they did not say, or showing a realistic scene that never happened. Using AI for a script, ideas or captions does not need that label (YouTube Help: Disclosing use of GenAI content). If you use an AI avatar or a cloned voice, tick the disclosure box during upload.

And one rule of your own: never let AI invent results, testimonials or numbers about your business. One made up claim can cost you more trust than a month of posts earns.

If you would rather have a tool draft the week for you, see how to build an AI social media content generator from one saved prompt and a short voice note.

Questions people ask

Which AI tool is best for content creation?

For writing, ChatGPT and Claude both do the job well. The difference is smaller than the difference between a good and a lazy process. Pick the one you already use, set up the voice file, and stick with it for a month before you judge.

Will people know I used AI?

They will notice generic writing, whether a person or a machine wrote it. If the piece has your own example, your opinion and your way of speaking, most readers simply see a useful post. That is why Steps 1, 2 and 5 matter more than the tool.

How much content should I make with AI?

As much as you can edit well. One strong main piece a week plus its smaller versions is a realistic start for most professionals. More only helps if the quality stays the same.

Can AI help me find ideas when I have nothing?

Yes, with a nudge. Ask: “Here is who I help and what I do. List 20 questions these people probably ask before they hire someone like me.” Then pick the ones you have actually heard from clients. Those are your strongest topics.

Is it okay to use AI content on my website?

Yes, if it is genuinely helpful and checked. Google’s own guidance focuses on whether content is useful, not on how it was made. Our guide on writing website copy with AI covers the pages themselves.

Start with one idea this week

You do not need a content calendar, a new tool or a whole weekend. You need one real idea from your work, ten minutes to talk it through, and the prompts above. Do it once this week. Next week it will already feel easier, because the voice file is set up and you know the steps.

If you want to learn more practical ways like this to let AI take the boring parts of your work, join our free Launchpad.

You open Gmail, bounce to a Doc, jump into a Meet recap, then stare at a spreadsheet someone dropped in chat. Every app has its own sparkle button, and you still write the first draft yourself because switching tools feels slower than just doing the work.

Here is the short answer on how to use Gemini for work. Keep Gemini inside the Google app you are already in. Use the side panel in Gmail to draft and catch up, in Docs to write and rewrite, in Sheets to build formulas and summaries, in Meet to capture notes, and in Chrome to work across the tabs you opened for a project. Save the prompts that work as a reusable skill, so you do not retype them every morning. Gemini does the first pass. You still decide what goes out the door.

This guide walks you through that desk system, with prompts you can copy today.

Want more short, practical AI lessons for work like this one? They all start on the RobotsHelp home page.

How to use Gemini for work inside the apps you already open

Gemini shows up in two places, and mixing them up is why a lot of people feel lost.

Inside Gmail, Docs, Sheets, Slides, Drive and Chat, look for Ask Gemini at the top right. That opens a side panel next to the file or inbox you have open. In Docs you can also switch the same helper to a bar at the bottom of the page (Google Workspace Learning Center).

The Gemini app at gemini.google.com, or the mobile app, is better for longer chats, research across several files, and the reusable instructions Google now calls skills. If the work already lives in a Google file or your inbox, stay there and open the side panel. If you need to compare sources or reuse a polished prompt, go to the Gemini app.

New to Gemini? Start with our Gemini tutorial for beginners, then come back here for the workplace setup. If you are still choosing an assistant, ChatGPT vs Claude vs Gemini compares the three. If your team lives in Claude, use how to use Claude for work as the matching guide.

What you need before you start

• An eligible account. Gemini inside Docs, Sheets, Slides and the other editors needs an eligible Google Workspace plan or a Google AI plan such as Google AI Plus, Pro or Ultra (Google Docs Editors Help).
• The right Gmail setup. People on an eligible Workspace or Google AI plan can use many Gmail features in more countries. On a personal Gmail account without a paid plan, some features, including Help me write and suggested replies, are limited to the United States (Gmail Help).
• Admin access, if this is a work account. Your administrator decides whether Gemini is on for your organization, and some meeting features need extra consent settings. If you cannot see Ask Gemini, ask them before you assume the product is broken.
• Your real files. Draft from the email thread, the brief, or the spreadsheet in front of you. That beats asking Gemini to invent the project from memory.

One privacy point is worth knowing up front if you handle client work. For businesses, education and public-sector customers on Google Workspace with Gemini, Google says your prompts and generated content stay inside your organization, and submissions are not reviewed by humans or used to train generative AI models. Your existing Workspace protections, such as data-regions policies, still apply (Google Workspace with Gemini FAQ). A personal Gmail chat is a different setup. Do not paste a client file into a personal Gemini window if your company asked you to keep that work in Workspace.

Start in Gmail: drafts, replies and catch-up

Most workdays start in the inbox, so start Gemini there too. Open Gmail on a computer, then click Ask Gemini at the top right. You can ask it to summarize a thread, list action items, find mail from one person last week, or catch you up on a project (Google Workspace Learning Center).

When you need to write, use Help me write on a new message or a reply. Say who the email is for, what you want them to do, and the tone you want, then edit the draft in your own voice. Proofreading can tighten grammar and tone after that, and suggested replies appear under some threads if your plan includes them (Gmail Help).

For a client email, this is a prompt that keeps Gemini from sounding generic:

Write a reply to this email.
Audience: [client or colleague name], [their role].
Goal: [the one thing I want them to do or know].
Tone: clear, warm, and direct. No filler and no exclamation marks.
Use only facts from this thread. If a date, price, or name is missing, ask me before you invent it.
Keep it under 160 words.
End with one clear next step and a suggested meeting window of [options].

Then do a 30-second check that Gemini cannot do for you: names, numbers, promises, and anything that would embarrass you if it were wrong. Change at least one sentence so it sounds like you.

For a deeper walk through drafts, summaries and inbox catch-up, use our full guide to Gemini in Gmail. The page you are reading stays on the wider work system. That one stays inside your inbox.

Draft and rewrite documents in Google Docs

Open the Doc you need to write, then open Ask Gemini. You can work from the side panel or switch to the bottom bar, which is handy when you want to edit a selected passage without losing sight of the page (Google Workspace Learning Center).

Gemini in Docs can draft or refine text, summarize a long file, generate an image, and match the writing style or layout of another Google Doc in Drive. That last part is the one people skip. If your team already has a proposal or status update that sounds right, point Gemini at that file and ask it to follow the same shape.

A project update works better with a tight prompt:

Draft a one-page project update using only this document and any files I mention with @.
Readers: [manager or client], who has 3 minutes.
Follow this structure:
1. What changed since [date]
2. What we shipped
3. What is blocked, and who can unblock it
4. The decision I need by [date]
Keep the tone factual. Do not add metrics I did not provide.
Flag anything in the source files that looks out of date.

Two habits make Docs drafts usable. Mention source files with @ so Gemini is not guessing from a blank page, and rewrite the opening paragraph yourself. Readers judge the whole note by those first lines, and that is where AI drafts usually sound canned.

If the document is research rather than an update, see ChatGPT vs Gemini for research before you pick a default tool for that job.

Get help with spreadsheets in Google Sheets

Sheets is where Gemini earns its keep for people who do not think of themselves as spreadsheet experts. Open your sheet, then open Ask Gemini and describe the table, formula or summary you need. Google’s own examples include building a tracker, summarizing a sheet, counting unique customers in a region, and writing a lookup formula (Google Workspace Learning Center).

Three built-in helpers are worth knowing by name:

• Enhanced Smart Fill looks at incomplete column pairs and predicts the rest, which helps when you are categorizing notes or standardizing names.
• The AI function can generate text, summarize cells or categorize a column from a prompt inside the sheet.
• Convert to table turns a loose block of data into a proper table so filters and formulas behave.

When a colleague dumps a raw export on you, try this:

This sheet is a raw export of [what it is] from [date range].
1. Tell me what each column appears to contain.
2. Point out blank cells, duplicates, and values that look like errors.
3. Propose a clean table with these columns: [list].
4. Write the formulas I need for [totals, unique counts, or a lookup], and explain each one in plain language.
5. Give me 5 summary bullets a manager could read in one minute.
Do not invent numbers that are not in the sheet.

Spot-check a formula on two or three rows you already understand. Gemini can write a tidy formula that still points at the wrong column. If the answer will go into a client report, compare the totals with a filter or a pivot table first.

Let Meet write the first notes, then you edit

If your meetings happen in Google Meet, you do not have to type a transcript while you talk. Take notes for me captures notes in a Google Doc and can share them with your team. If you join late, Summary so far helps you catch up. After the call, the organizer gets an email with a link to the recap, and the same link appears on the Calendar event (Google Meet Help).

A few limits matter. The feature needs an eligible Workspace edition or Google AI plan. It works in English, French, German, Italian, Japanese, Korean, Portuguese and Spanish, one spoken language at a time. Google recommends it for meetings from 15 minutes up to 8 hours. If host management is on, only hosts can turn notes on or off, and your admin may require everyone to consent first.

Treat those notes as a first draft. Fix names and decisions, and add anything agreed in a side chat. For the full setup, use our guide to Gemini for meeting notes. If you send the same recap after every standup, Gemini for automated status updates shows how to turn notes into a repeatable report.

How to use Google Gemini to organize tabs in Chrome for work

A work research session often ends with 20 tabs and no clear next step. Gemini in Chrome is built for that mess. It uses the page you are looking at, and on a computer you can share up to 10 open tabs so Gemini can compare or combine what is on them. You add a tab by typing @ and picking it, or by using Add tabs in the chat box (Google Chrome Help).

How to start:

1. Open Chrome on a computer and sign in. Gemini in Chrome does not work in Incognito.
2. Click Ask Gemini at the top of the browser. The first time, follow the prompts to opt in.
3. Share the tabs that belong to this task, up to 10, and leave unrelated tabs out.
4. Ask Gemini to compare them, extract decisions, or draft something you can paste into a Doc.

A prompt that turns a pile of tabs into a briefing:

I shared up to 10 tabs about [project or decision].
Write a briefing I can send to my manager.
Include:
- The question we are trying to answer
- What the sources agree on
- Where they disagree, with the source title next to each point
- Risks or costs mentioned in the pages
- A recommended next step, with one backup option
Quote or paraphrase only from the shared tabs. If something is missing, say what tab I should open next.
Do not invent vendors, prices, or dates.

Google is still rolling Gemini in Chrome out, so you may not see it yet. You need the current version of Chrome on a Chromebook Plus, Mac or Windows computer, a supported region and language, and a sign-in. If you use a work or school account and you are over 18, your administrator has to allow it. When you share a Workspace page such as a Google Doc, Gemini can read the full file rather than only what is visible in the tab.

Gemini helps you make sense of the tabs. Chrome’s own tab groups still help you separate “client A”, “hiring” and “invoice follow-up” so the next session starts cleaner. Use both.

Save the prompts that work as a skill

The prompts above only pay off if you stop rewriting them from scratch. In the Gemini app, reusable instructions are moving from Gems to skills. A skill is a saved set of instructions you can add to any chat by typing / and the skill name. Gemini can also apply a relevant skill on its own, and you can stack more than one skill in the same conversation (Gemini Apps Help).

Timing depends on your account. For personal Google accounts, if you are 18 or over, Gems become read-only on 17 November 2026 and are removed on 17 December 2026. Google says it will recreate your Gems as skills automatically. Workspace business, enterprise and non-profit accounts keep Gems until March 2027, and education accounts until June 2027. Skills are already available in Gemini chat for people over 18 on a personal account. Google says work and school access is coming.

To create a skill now in the Gemini app, open the sidebar, go to Customize, then Skills, and choose Create and Create manually. Give it a clear name, a short description of when to use it, and the instructions you would otherwise paste every time.

Here is a solid first skill for outbound work email:

Name: work-email-editor
When to use: any time I ask you to draft, rewrite, or tighten a work email.
Instructions:
You are my email editor for client and colleague messages.
Write in short sentences. Sound like a calm professional, not a marketer.
Never invent dates, prices, names, or commitments.
If a detail is missing, ask me one question instead of filling the gap.
Default structure: context in one line, the request, then one next step.
Keep most emails under 160 words.
If I paste a messy draft, return an edited version and a 3-bullet explanation of what you changed.

Make one skill per repeating job, not one giant “work” skill. A status-update skill, a meeting-agenda skill and an email-editor skill will stay sharper than a single assistant that tries to do everything.

A sample day at your desk

Here is how the pieces fit on a normal Tuesday.

Morning, Gmail then Docs. Open Ask Gemini and say, “Catch me up on the Horizon rollout emails since Monday. List anything waiting on me.” Use Help me write on the two replies that need a human. Then open yesterday’s brief, mention it with @, and ask for a one-page update in your team’s usual format. Rewrite the first paragraph before you send it.

Midday, Chrome then Meet. Share your vendor pages, a pricing FAQ and your brief with Gemini in Chrome, and ask for a comparison table plus the questions you still need to ask. Before the vendor call, turn on Take notes for me. Afterward, clean the Doc: fix names, mark the real decision, delete the small talk.

Late afternoon, Sheets. Open the numbers the vendor sent. Ask Gemini to explain the columns, write the totals, and give you five bullets. Check two rows by hand, then drop the bullets into your update Doc.

That is a full loop: inbox, document, research tabs, meeting notes, numbers. None of it required you to copy the same text into four different chat windows.

Photos are another everyday job Gemini can take off your plate. See how to use Gemini to edit photos for quick fixes to product shots and profile pictures.

What Gemini still gets wrong at work

Google is clear that Gemini can suggest inaccurate or inappropriate information, and that you should not treat it as legal, financial, medical or other professional advice (Google Workspace Learning Center). That is not fine print. It is the rule for anything you send to a client or a manager.

Watch for these failure modes in particular:

• Confident gaps. If you did not give Gemini a date or a price, it may still write one that sounds plausible. Ask it to flag missing facts instead.
• Wrong names and pronouns. Summaries mix up who promised what, especially in long email threads.
• Formulas that point at the wrong range. A tidy-looking Sheets formula is still wrong if column D is notes, not revenue.
• Meeting notes that miss the decision. “Take notes for me” hears what was said. It does not know which offhand comment was the actual yes.
• Tone that is too smooth. Client email should sound like you. If every message starts with “I hope this message finds you well,” delete that line every time.

If you would not put your name under the output after a two-minute review, it is not ready. Gemini is a fast first pass. You are still the person who sends it.

Once this system feels natural, keep going with short, practical lessons on the RobotsHelp home page.

Images are one more job to hand over: read how to use Gemini to create images for work.

If you want a structured path to build these habits, the Gemini AI course guide compares the free Google lessons and lays out a 2 week plan.

You finally booked the certification exam. Or your manager asked you to “get up to speed” on a new tool by the end of the month. The material sits in a folder of PDFs, and your study time is forty minutes on a weeknight, after everything else.

Here is the short answer on how to use Gemini to study. Put your material in one Gemini study notebook, take its diagnostic quiz to find your gaps, use Guided Learning for the topics you do not understand yet, turn each chapter into quizzes and flashcards you answer from memory, and listen to an Audio Overview on your commute. Then come back two or three days later and test yourself again. Gemini does the setup work. You still do the remembering, and that is the part that makes it stick.

This guide walks you through that system step by step, with prompts you can copy tonight.

Want more short, practical AI lessons for work like this one? They all start on the RobotsHelp home page.

What “using Gemini to study” should actually mean

The trap is easy to fall into. You upload a 200-page guide, ask for a summary, read the summary, and feel like you learned something. Most of that feeling fades by Friday.

Learning research is clear on two habits that work better than rereading. The first is testing yourself: trying to pull an answer out of memory beats reading it again for long-term recall (Roediger and Karpicke, 2006). The second is spacing: spreading your practice over several days beats cramming it into one sitting (Cepeda and colleagues, 2006).

So the job you give Gemini is not “explain this to me.” It is “build me the questions, check my answers, and show me what I keep missing.” Every step below follows that rule.

What you need before you start

• A Google account signed in at the Gemini website or app. The free plan works for most of this guide.
• Your real material. The official exam guide, course slides, your own notes, a product manual. Studying from your actual source beats studying from whatever the AI remembers about the topic.
• Keep Activity turned on if you want study notebooks. Google’s help page says study notebooks, notebook memory and adding Drive files all need it (Gemini Apps Help: notebooks).

One limit worth knowing up front. Without a paid plan, Gemini can hold about 32,000 tokens of context at once. Google AI Plus raises that to 128,000, and AI Pro and Ultra to 1 million. Google’s own rule of thumb is that 1 million tokens is about 1,500 pages (Gemini Apps Help: limits). That works out to roughly 50 pages on the free plan. If your guide is longer, feed it one chapter at a time. That is a better way to study anyway.

New to Gemini in general? Our Gemini tutorial for beginners covers the basics first. If you are a student, check if you qualify for the free upgrade in Gemini for students free before you pay for anything.

Step 1: Build a study notebook for one goal

A notebook is a dedicated space in Gemini that remembers your sources, your instructions and your past chats about one project. A study notebook is a special type built for learning. According to Google, it gives you a diagnostic quiz to find your gaps, builds short interactive lessons from the results, and updates your progress as you complete activities (Gemini Apps Help).

How to set one up on a computer:

1. Go to the Gemini website and open the sidebar.
2. Under Notebooks, click New notebook.
3. Click Study and learn.
4. Tell Gemini your goal in one sentence, with a date. For example: “Help me pass the PMP exam on 15 January. I work full time and can study 40 minutes on weeknights.”
5. Click Add files and upload your material from your device or Google Drive.

Two things to know. You cannot edit the goal of a study notebook later, so make it specific. And Google says study notebooks are on the Gemini website, not yet in the Gemini app on Mac. If you do not see Study and learn yet, a regular notebook with the instructions from Step 2 does most of the same job.

Make one notebook per exam or skill. Not one giant “learning” notebook. When a notebook only knows about your Excel certification, its quizzes stay on your Excel certification.

Step 2: Give Gemini tutor rules

Out of the box, Gemini tends to answer your question and move on. For studying, you want the opposite: a tutor that makes you work for it. In any notebook you can add standing instructions under More, then Notebook settings. If you prefer, put the same text into a Gem, a reusable custom version of Gemini with its own instructions and knowledge files (Gemini Apps Help: Gems).

Copy this and adjust the brackets:

You are my study coach for [exam or skill]. My exam date is [date].
Use my uploaded sources first. If something is not in my sources, say so before you answer from general knowledge.
Do not give me the answer straight away. Ask me one question at a time.
If I am stuck, give a small hint, not the solution.
When my answer is vague or wrong, tell me exactly which part and why.
Keep a running list of the topics I get wrong. At the start of each session, quiz me on two of them first.
Keep explanations short and use examples from [your job or industry].

The last line matters more than it looks. “Explain earned value with an example from a marketing agency” lands far better than a textbook example about building bridges, if you work at a marketing agency.

Step 3: Use Guided Learning for the parts you do not get yet

Some topics need teaching before testing. That is what Guided Learning is for. It is a mode in the Gemini app that breaks a topic into steps and asks you questions along the way instead of handing you the full answer. Google describes it as a tool that “coaches students through complex concepts, practice quizzes, and custom visual aids” (Google, November 2026).

To turn it on at the Gemini website, click Add files in the text box, then More tools, then Guided Learning. On the phone app it sits at the bottom of the same menu. You can upload a file or a photo of a page along with your question (Gemini Apps Help: learning tools).

Keep your question narrow. “Teach me statistics” gives you a lecture. This gives you a lesson:

I keep mixing up [concept A] and [concept B] in chapter 4 of the file I uploaded.
Walk me through the difference step by step. Check that I understand each step before you move on.
Finish with three questions where I have to pick which one applies.

Is there evidence it helps? Some. In a randomized trial with more than 300 University of Tokyo students, Google reported that students with Guided Learning improved their test scores by 8.1 percentage points in a half-day statistics session, against 5.3 points for students using standard web tools (Google). Two honest caveats: Google ran the study on its own product, and the group was university students, not working adults. Treat it as a good sign, not a promise.

Step 4: Turn every chapter into a quiz and flashcards

This is the heart of the system. Gemini can build interactive quizzes, flashcard decks, practice tests and study guides that open in the Canvas panel next to your chat. You start your prompt with what you want, then add the material (Gemini Apps Help: quizzes and flashcards):

• “Create a quiz about…”
• “Create flashcards about…”
• “Create a study guide about…”

In a quiz you can ask for a hint, ask follow-up questions in the chat beside it, and click More questions when you finish. Wait a moment at the end and Gemini shares your strengths and focus areas. Flashcards can be shuffled and read aloud.

A prompt that gets you better questions than the default:

Create a quiz about chapter [number] using only the file I uploaded.
10 questions. Mix multiple choice with short scenario questions like the ones on the real exam.
Make at least three questions about the ideas people most often confuse.
After each answer, show which page or section of my file the answer comes from.

That last line is your safety net. If Gemini cannot point to where an answer came from, check it before you learn it.

One rule for flashcards: answer out loud or on paper before you flip the card. Flipping and nodding is rereading with extra steps. If you want to build quizzes for other people, such as a team training, see how to make a quiz with AI.

If your sign-in is a work account

Google notes that on a work or school Google account you cannot share practice tests, quizzes, flashcards or study guides (Gemini Apps Help). You can still use them yourself. Also check your company’s rules before you upload internal documents to any AI tool.

Step 5: Listen to your material with Audio Overviews

You have dead time every day. The commute, the walk to lunch, folding laundry. Audio Overviews turn a document or slide deck into a podcast-style conversation between two AI hosts. Upload the file, ask Gemini to create an Audio Overview, and Google says it usually takes about 3 to 5 minutes to generate. You can download it from the player (Gemini Apps Help: Audio Overviews). Google also notes that the audio itself is only made in English, and the feature is not available to users under 18.

Use it for orientation and review, not as your only source. An AI summary can skip the exceptions and footnotes, which is exactly where exam questions like to hide. A good pattern: listen to the overview of a chapter on Monday morning, then take the quiz on that chapter Monday night.

Step 6: Space it out across the week

Now put it on the calendar. Here is a simple week for someone with 40 minutes on weeknights and an hour on Saturday:

• Monday: Audio Overview of chapter 3 on the commute. Evening: Guided Learning on the one concept that confused you, then a 10-question quiz.
• Tuesday: Flashcards from chapter 3. Then the two weak topics from your notebook’s running list.
• Wednesday: Start chapter 4 the same way as Monday.
• Thursday: Quiz on chapter 3 again, without looking at your notes first. This is the spaced repeat.
• Saturday: Mixed quiz across everything so far. Then the “explain it back” test below.

If planning the week is the hard part, ask Gemini for help with that too. Our guide on how to use AI to plan your day shows how to fit study blocks around real work.

The explain-it-back test

Once a week, close your files and type out everything you remember about one topic, as if you were teaching a new colleague. Then paste it in:

Here is my explanation of [topic], written from memory: [paste].
Compare it with my uploaded sources.
List what I got right, what I got wrong, and what I left out, in that order.
Then give me three new questions on the parts I got wrong.

It feels uncomfortable. That is the point. The gaps it shows you are the gaps the exam would have found.

A real-world example: learning a new tool for work

Picture Jordan, an account manager whose company is moving its reporting into Looker Studio. There is no exam, but there is a deadline: build the first client dashboard in three weeks.

• Jordan creates a study notebook with the goal “Build a client dashboard in Looker Studio by 1 December,” and adds the company’s internal how-to doc and two pages of the official help as website sources.
• The diagnostic quiz shows the weak spot is data sources and blending, not charts. So that is where the week goes.
• Guided Learning walks through blending with examples from client reporting, because the notebook instructions say so.
• Two short quizzes a week, plus one explain-it-back on Saturday.

Jordan could have watched hours of video. Instead, the questions point straight at what Jordan does not know yet, and the rest gets skipped. The same system works for a language, a sales methodology or an internal compliance course.

Mistakes that waste your study time

• Studying from Gemini’s memory instead of your material. Always upload the official source and ask Gemini to cite it.
• Letting it answer for you. If you ask “what is the answer to question 7,” you skipped the part that builds memory.
• One endless chat. Long chats use more of your usage limit, because Gemini re-reads the history each time. Google’s own tip is to start a new chat when you move to a new topic (Gemini Apps Help). A notebook keeps your sources and context, so new chats inside it are cheap.
• Summaries as notes. A summary is fine as a map. Your notes should be the questions you got wrong and the correct answers in your own words. For turning dense PDFs into notes you will actually reuse, see how to turn a PDF into notes.
• Trusting every answer. AI tools get things wrong with total confidence. For anything that will be on the exam, check the page it cited.

Gemini or another tool for studying?

Gemini’s strength is that the study tools are already built: quizzes, flashcards, study notebooks and Audio Overviews with no setup. If you live in Google Drive at work, your material is already one step away. If you mostly study by reading long, dense documents and arguing with them, Claude is also strong. We compare both in Claude vs Gemini for studying, and the Claude version of this routine is in how to use Claude to study. The best tool is the one you open again tomorrow night.

FAQ

Can I use Gemini to study for free?

Yes, for most of this guide. Quizzes, flashcards and Guided Learning work in the Gemini app, and the free plan has standard usage limits. The free context window is smaller (32,000 tokens), so upload one chapter at a time. Feature availability can differ per plan, so check Google’s limits page for your account.

Why don’t I see study notebooks or Guided Learning?

Google is rolling these out gradually, especially in the mobile app. Try the Gemini website first. For study notebooks, check that Keep Activity is on. In the European Economic Area, you may need to connect Gemini Notebook yourself under Connected Apps (Gemini Apps Help).

Can Gemini make a practice test for my certification?

It can generate a practice test from your uploaded material with “Create a quiz about…” or a full-length prompt. Google lists a few supported standardized tests by name, such as the SAT. For professional certifications, upload the official exam outline so the questions match what is actually tested, and treat the questions as practice, not a copy of the real exam.

Is it safe to upload work documents?

Check your company’s AI policy first. Some companies only allow Gemini through a work Google account managed by IT, and some do not allow internal files in AI tools at all. When in doubt, study from the public exam guide or official help pages instead.

How long should a Gemini study session be?

Short and frequent beats long and rare, because spacing is what helps memory. Thirty to forty-five minutes with at least one self-test in it is a good default. Stop while you still remember what you got wrong, and start the next session with those topics.

Sources

• Gemini Apps Help: Organize your projects with notebooks (study notebooks)
• Gemini Apps Help: Use learning tools (Guided Learning)
• Gemini Apps Help: Create quizzes, flashcards, practice tests and more
• Gemini Apps Help: Generate Audio Overviews
• Gemini Apps Help: Limits and feature availability
• Gemini Apps Help: Use Gems
• Google: New research shows how Guided Learning supports students (November 2026)
• Roediger and Karpicke (2006), Test-enhanced learning, Psychological Science
• Cepeda et al. (2006), Distributed practice in verbal recall tasks, Psychological Bulletin

Your next step

Tonight, pick one chapter you have to learn anyway. Make a study notebook, upload it, and take one quiz without looking. That is your whole first session. If you want a new practical AI lesson for your work every week, join here:

The call just ended. You have a page of half sentences, two names you are not sure about, and a nagging feeling someone agreed to do something by Friday. You were too busy listening to write it all down. That is normal, and it is exactly the job Gemini can take off your plate.

Here is the short answer. If your Google account comes with Gemini in Google Meet (an eligible Google Workspace plan or Google AI plan, on the organizer’s account), click Take notes for me at the top right of the call, then Start taking notes. After the meeting, Gemini saves a Google Doc with a summary, decisions and next steps to the organizer’s Drive and attaches it to the Calendar event. If you do not have that, you can still use Gemini for meeting notes: paste your rough notes or a transcript into Gemini in Google Docs or the Gemini app and ask for decisions and action items with owners. Both paths are below, step by step.

Want a free starter path for using AI like this in your daily work? Start on the RobotsHelp home page.

Two ways to use Gemini for meeting notes

People mix these up, so it helps to see them side by side. One listens to the meeting. The other works on text you already have.

If most of your calls already happen in Google Meet, start with the first one. If you live in Zoom or Teams, skip to the Docs section. The prompts there work for any meeting.

Before you start: three things that block the button

Nothing is more annoying than looking for a button that is not there. Google’s own help page lists the usual reasons (Google Meet Help: Take notes for me):

• The organizer’s plan does not include it. It is the organizer’s Workspace edition or Google AI plan that counts, not yours. If a client sends the invite from a plan without it, you will not see the option in their meeting.
• Your admin has not turned it on. On a work or school account, the admin decides whether the feature is available.
• Smart features are off in your Google account settings.

Two more practical limits. Google recommends meetings of 15 minutes up to 8 hours, and the meeting needs to be spoken in one supported language: English, French, German, Italian, Japanese, Korean, Portuguese or Spanish. A bilingual call is not supported yet, so if your team switches between Dutch and English halfway, expect patchy notes.

How to turn on Take notes for me in Google Meet

You can switch it on during the call, before the call in Calendar, or once for every meeting you host.

During a call

1. Join the meeting on your computer at meet.google.com.
2. At the top right, click Take notes for me (the pencil icon).
3. Optional: open Settings to choose who gets the notes, the notes length (Standard or Longer) and the meeting language.
4. Click Start taking notes. Everyone in the call sees the pencil icon, so they know notes are running.

Before a call, from Google Calendar

1. Open the event in Google Calendar and click Edit event.
2. Make sure Google Meet is added, then open Video call options.
3. Click Meeting records and tick Take notes with Gemini.

This is the version I recommend for recurring meetings. You set it once on the weekly team call and stop relying on memory. Inside Meet there is also an option to keep it on for future instances of the same meeting.

For every meeting you host

In the Meet settings (gear icon), under meeting records, you can choose to start notes automatically when you host. Since late September 2026 there is also an option for meetings with three or more guests. Google says that on Business Standard and Business Plus plans, the admin default for that 3+ setting is on, so notes may start by themselves in bigger meetings unless you or your admin change it (Google Workspace Updates, September 2026). Worth a quick look before your next sensitive client call.

Where the notes land, and what is in them

Shortly after the meeting ends, Gemini creates a Google Doc in the organizer’s Drive, inside a “Google Meet” folder with a sub-folder per meeting. The doc is attached to the Calendar event, and the organizer gets an email with a link. Depending on the sharing setting, invited guests get access too.

The doc has two tabs:

• Quick notes: the short version with the main takeaways and action items. This is the default tab.
• Full notes: a longer overview with a summary, decisions, next steps and details.

The decisions section tags each outcome as Aligned, Needs Further Discussion, Disagreed or Shelved. The next steps section lists the action items and who they were assigned to, when that was said out loud. That last part matters, and we will come back to it.

Who should get the notes?

Hosts and co-hosts can pick one of three options under “Send notes to”: all invited guests including people outside your organization, only invited guests inside your organization, or hosts and co-hosts only. A simple rule of thumb: internal team meetings go to everyone invited, client calls go to hosts only until you have read the doc. One detail that trips people up: “invited guests” means everyone on the Calendar invite, not who actually showed up.

How to use Gemini to create action items from meeting notes

The Next steps section is a good start. It is rarely good enough to forward as is. Here is the five-minute routine I use right after a call, while it is still fresh.

1. Open the notes doc from the Calendar event or the email.
2. Click Ask Gemini at the top right of Google Docs. The side panel opens with the doc already in view.
3. Paste this prompt:

From these meeting notes, make a table of action items with four columns: Task, Owner, Due date, Source line.
Rules:
- Only include tasks that were actually agreed in the meeting.
- If no owner was named, write "Owner?" instead of guessing.
- If no date was named, write "No date" instead of guessing.
- Quote the line from the notes the task comes from.
After the table, list open questions nobody answered.

4. Fix the gaps. Every “Owner?” is a message you send today, not a task that quietly dies.
5. Insert the table into the doc, so it lives next to the notes instead of in a chat history you lose when you close the tab.

Why the “Source line” column? Because it makes checking fast. If a task has no line behind it, Gemini guessed, and you delete it. Google itself says Gemini output can be incomplete or wrong, so a quick human pass is part of the job, not an extra.

The Docs side panel can also pull in other files. Type @ in the prompt and pick, for example, your project tracker, then ask: “Update the task list in @Project tracker with the new action items from this meeting.” Google describes this “reference other files” option in its help page for Docs (Collaborate with Gemini in Google Docs).

How to use Gemini to summarize meeting notes in Google Docs

No Take notes for me? Your meeting ran on Zoom? You can still use Gemini on your own notes or an exported transcript.

1. Paste your rough notes or the transcript into a new Google Doc.
2. Click Ask Gemini. When a doc already has text, the side panel shows a summary of it.
3. Ask for the shape you need. To summarize only one part, highlight that text first and ask Gemini to summarize the selection.

Three prompts that cover most meetings:

1. Summary for people who missed it:
"Summarize this meeting in 5 bullets for someone who was not there. Start with the decisions. Keep each bullet under 20 words."

2. Client recap:
"Write a short recap email to the client based on these notes. Include what we agreed, what we need from them, and the next meeting date if it was mentioned. Friendly, plain English, no jargon."

3. Decision log:
"List every decision in these notes, one line each, with the reason given in the meeting. If a reason was not mentioned, say so."

Without a Workspace or Google AI plan, the free Gemini app at gemini.google.com takes the same prompts. Paste the text, ask, copy the answer back. It is one extra step, but the result is the same kind of summary. Do check your company’s rules first before you paste client details into any AI tool.

If your team works in ChatGPT instead, the same idea works there too. Our guide on how to use ChatGPT to summarize meeting notes has a prompt set built for that.

Use Gemini to automate meeting notes and action items

“Automate” here does not mean nobody touches anything. It means the boring parts happen by default and you only do the five minutes of judgment. A setup that works for a small team:

1. Turn notes on by default for the meetings you host (Meet settings), or per recurring event in Calendar.
2. Choose the sharing setting once, for example invited guests in your organization, so nobody has to ask “can you send the notes?”
3. Keep one follow-up doc per project and use the @ reference prompt above to add new action items to it after each meeting.
4. Draft the follow-up email from the notes doc with Ask Gemini (“Draft an email to the team with the action items and owners from this doc”), then read it and send it yourself.

Want the weekly version of this, where Friday’s update writes itself from a Sheet? That is covered in Gemini for automated status updates. And for a drag-and-drop agent that preps you before meetings, see our n8n AI agent workflow guide.

Run the meeting so the notes come out right

Gemini can only write down what was said. Most “bad AI notes” are really meetings where nobody said the important part out loud. A few habits fix most of it:

• Say the owner and the date. “Sam, can you send the quote by Thursday?” lands as a clean next step. “Someone should look at the quote” does not.
• Close each topic with one sentence. “So we go with option B.” That is what ends up in the decisions section.
• Start notes at the start. Notes only cover the part of the meeting after they were switched on.
• Spell out names and acronyms once. Internal product names and client names are where transcripts go wrong most often.
• Use an agenda. A clear order of topics gives clearer sections in the doc. Our guide on how to use AI to create a meeting agenda takes about ten minutes.

Picture a supplier call that ends with “follow up on pricing” scribbled in a notebook, and a week later nobody knows who promised what. Now end the same call with “So: I send the specs today, you send the new price by Wednesday.” That sentence has two owners and two dates, which is exactly what the next steps section needs, and your recap email is half written.

Privacy and etiquette: tell people before you start

Google Meet shows everyone that notes are being taken, and Google asks the person who starts it to let the group know. Do it anyway, out loud. A short “I am letting Gemini take notes so I can focus, is that okay?” costs five seconds and keeps trust intact.

• Your admin can require participants to give consent before note-taking, recording or transcription starts.
• Internal participants can stop the notes at any time, for example before a confidential topic. When host controls are on, only hosts and co-hosts can.
• The notes doc follows your organization’s Meet retention policy.

For client meetings, a sensible default is to share notes with hosts only, check them, then send a clean recap. You stay in control of what leaves your company.

Gemini or a separate AI note taker?

If your meetings run on Google Meet and your company already pays for a Workspace plan that includes it, Gemini is the easy choice: nothing to install, no bot joining as an extra guest, and the notes sit in Drive next to your other files. It is limited to Meet calls (and in-person use), one language per meeting, and you cannot plug it straight into a CRM.

If you jump between Zoom, Teams and Meet, or you need CRM sync, a dedicated tool fits better. We compared the main options honestly in best AI note taker for meetings. And if your meeting notes need to become a proper report for your boss or client, see turn meeting notes into a report people can act on.

Meeting notes are one part of the workday. If you also want Gemini in Gmail, Docs, Sheets and Chrome, our guide on how to use Gemini for work walks through a daily system across those apps, with copy-ready prompts you can reuse.

FAQ

Is Gemini for meeting notes free?

Take notes for me in Google Meet is not part of a free personal account. It needs an eligible Google Workspace edition or Google AI plan on the organizer’s account. Summarizing notes you paste into the free Gemini app does work without a paid plan.

Can Gemini take notes in Zoom or Microsoft Teams?

Take notes for me works in Google Meet. For Zoom or Teams, export the transcript or your own notes and use Gemini in Docs or the Gemini app with the prompts above.

Why don’t I see the Take notes for me button?

Most often the organizer’s plan does not include it, your admin has not enabled it, or smart features are turned off in your account. Google lists these three in its troubleshooting section.

Does Gemini assign action items to the right people?

It assigns them when the meeting said who does what. If nobody named an owner, there is nothing to assign. That is why the prompt above writes “Owner?” instead of guessing, and why a quick read before you share is worth it.

Can I edit the notes before others see them?

Yes. Set “Send notes to” to hosts and co-hosts only, edit the doc, then share it or send a recap yourself.

Try it on your next meeting

Pick one recurring meeting this week. Turn on Take notes for me in the Calendar event, or paste your notes into Gemini afterwards. Run the action item prompt, fix the “Owner?” lines, and send the recap within the hour. That single habit saves most people the worst part of meetings: the “wait, who was doing that?” message three days later.

Want more short, practical routines like this one?

Using another tool? The same idea works with our AI meeting minutes generator prompt in ChatGPT or Claude.

If your calls run in Teams instead, see Otter AI for Microsoft Teams for how to let the notetaker in and when Copilot recaps are the better pick.

You heard there is a way to get Gemini for students free for a whole year. Then you opened three tabs, saw a payment form, and wondered if this is a real offer or a trial that quietly bills you later. You are not being dramatic. The offer is real, and the billing details matter.

Here is the short answer. If you are 18 or older and enrolled at a qualifying higher education institution, Google is offering eligible students 12 months of a paid Gemini plan at no charge, as long as you redeem by 31 December 2026. In the United States that plan is Google AI Pro (listed at $19.99 a month). In many other countries it is Google AI Plus. You claim it on a personal Gmail account, verify your student status with SheerID, and add a payment method. If you do nothing at the end of the year, Google says the plan converts to a normal paid subscription. Cancel before the year ends if you do not want that charge.

Want a free starter path for using Gemini and other AI tools in real work and study? Start on the RobotsHelp home page.

What Gemini for students free actually includes

Free Gemini still exists for anyone with a Google account. The student offer is different. It unlocks the higher paid tier for a year, which means more usage room and the study extras that run out fast on the free plan. Google announced the current round on 19 August 2026, with new study tools alongside the 12-month plans (Google blog).

Features shift as Google updates its plans, so treat the table as a snapshot and check the live pages before you decide. The student page lists the US extras as higher usage limits, 5 TB of storage, Gemini Spark and Gemini Omni (Gemini for Students). Google describes Omni as a model that can work with text, images and video (subscription update), and Spark as an assistant that can take actions across Google products (fall 2026 plan updates).

If you are comparing paid plans for work rather than claiming a student year, see our guide to ChatGPT vs Gemini Pro. Google also offers a separate student bundle of Google AI Pro plus YouTube Premium at a discount in select markets. That is not the same as the free 12-month plan.

Who qualifies, and who gets turned down

Google’s help page and offer terms are stricter than “any student with an email.” In plain language, you need all of the following (Google One Help: claim a student offer; offer terms):

• You are 18 or older. Your age must be verified on your Google Account, and the birthday you enter for student verification must match that account.
• You are enrolled at a higher education institution in a country where the plan in your offer is supported.
• You verify that status through SheerID, the third-party checker Google uses.
• You use a personal Google Account. Google says that means a personal Gmail address. University or school-issued accounts are not eligible, and the offer is not available on a school-issued Workspace for Education account.
• You have a Google Payments account with a qualifying payment method at sign-up. The year is free. The card is still required.

People who usually cannot claim it:

• Anyone under 18, or anyone whose Google Account is supervised through Family Link.
• Students at secondary school only. Google asks for a higher education institution, not a high-school enrollment.
• Members of a shared Google One family plan who did not create the plan, unless they leave that family group first or use a different personal Google Account.
• People whose current Google One membership comes through a third-party partner, Pixel Pass or Google Fi. Those partner plans usually need to be cancelled before the student offer will go through.
• Students in places Google excludes. The general help page excludes Hong Kong, Macau and Tunisia. The international Plus offer also excludes Bolivia, Albania and Canada, among others listed in Google’s footnotes. If you study in Canada, do not assume the free year applies. Check the official offer page for your country.

Good news if you already used last year’s trial: the Gemini student page says both new users and students whose 2025 AI Pro trial has expired can qualify. Eligibility has to be verified again each year to keep access.

How to claim the offer this week

Set aside about 20 minutes, plus waiting time if SheerID needs to review a document. Do this on a laptop if your phone keeps dropping campus Wi-Fi. Google’s own troubleshooting notes that a weak connection can time out the page.

1. Use the right Google Account. Sign in to the personal Gmail account you want to keep for the next year. Student verification cannot be moved to another account later. If you verify on account A and try to finish sign-up on account B, the claim fails.
2. Confirm your age on the account. Open Google’s age verification page at myaccount.google.com/age-verification and make sure it matches the birthday you will type into SheerID.
3. Start on the official offer page. Go to one.google.com/ai-student. Do not begin from a random forwarded link. Google tells people who hit errors to start there, and to try another browser if needed (Chrome, Firefox or Safari).
4. Select the student offer that matches your country. US students should see the Google AI Pro year. Students in supported countries outside the US should see Google AI Plus. Read the plan name before you continue.
5. Verify with SheerID when asked. Many people get approved from school records SheerID already knows. If you are asked for proof, upload a document that clearly shows your full first and last name (matching the form), your university name or official logo, and a date that shows current enrollment. A student ID with only initials, or a card with no date, is a common reason for rejection. Google suggests a tuition receipt, class schedule or student-portal screenshot for the current academic year, and you can upload more than one file.
6. Finish the subscription flow and add a payment method if you do not already have one on the account.

SheerID review often takes a few minutes. In busy back-to-school periods it can take up to a few days. Do not resubmit the form over and over. Attempts are limited, and after three failed document checks you need SheerID support rather than another upload on the same account. If you see “verification limit exceeded,” Google says those student details were already used on another Google Account. Each student can redeem the offer once.

Other blockers Google lists: your Google Play or Payments country does not match where you live and study, you are on a family plan you did not create, or you are still billed through Apple, a carrier or another partner. If the page fails on your phone, try a private browsing window or a computer.

Your first hour after you get in

Do not spend the first hour generating a stack of unused notes. Use the extra limits on one course you actually have this week.

1. Open the student hub. Google built a hub for student tools at gemini.google.com/students. From there you can start a study notebook, build flashcards or take a practice quiz. The hub is rolling out on mobile and web for signed-in consumer accounts.
2. Create one study notebook from real class material. Upload a syllabus, lecture slides or your own notes for a single course. Google says Gemini then builds a plan, and you can take a diagnostic quiz so the lessons focus on gaps instead of chapters you already know. With a paid AI plan you can create more notebooks with more sources than on the free tier.
3. Put key dates on your calendar, with permission. Google says Gemini can schedule exam and assignment dates from a syllabus into Google Calendar when you allow it. That is useful. Still read the syllabus yourself. A missed date in an AI parse is your problem, not Gemini’s.
4. Try one visual explanation. Interactive visualizations can turn a concept into a table, a grid or a 3D simulation you can rotate. Google’s examples include asking to see how DNA works in 3D, or watching a pendulum trade energy. Use this for something you are actually stuck on, not as a toy.
5. Talk one problem through in Gemini Live. You can speak through a problem set or files from Drive. Google is also bringing Deep Research into Gemini Live, so you can ask for a longer report, leave the chat, and get a notification when it is ready. Check every source before you cite it. We cover that research mode in ChatGPT vs Gemini for research.

A simple study routine that uses the extra room

Higher limits only help if you study with your own material. Gemini’s student page is clear about the point: the product is meant to help you understand, not just generate answers, and prompts that need course materials work better after you upload the files.

Here is a routine you can run in about 45 minutes for one upcoming class or exam:

1. Upload this week’s slides or notes into a notebook for that course.
2. Ask for a short diagnostic quiz, then write your answers before you look at explanations.
3. Pick the two questions you missed and ask Gemini to walk through the idea step by step, using only the uploaded material.
4. Close the chat and explain the same idea out loud, or in a fresh note, without looking. If you cannot do that, you did not learn it yet.
5. Ask for five new quiz questions on those weak spots only.

A prompt that stays honest looks like this:

I uploaded my own class notes and slides for [course name].
Quiz me on [topic] in 8 short questions.
After each answer I type, tell me if I am right or wrong
and point to the part of my notes that supports the correct answer.
If my notes do not cover it, say "not in my notes" instead of
filling the gap from elsewhere.
Then give me a 5-bullet study plan for what I still need to review.

If you want a study method that is built around active recall (and you are happy to work in Claude as well as Gemini), use our guide on how to use Claude to study. If you are still choosing a model for learning, read Claude vs Gemini for studying. This page stays on how to get the free student year and what that year unlocks.

A word on assignments. Check your course rules before you upload materials or hand in anything an AI helped produce. Many courses allow AI for practice and explanations, and forbid submitting generated essays as your own work. Using Gemini to quiz you, explain a concept, or critique a draft you wrote is a different thing from asking it to write the paper. When you want feedback, paste your draft and the rubric, and ask what is unclear. Keep the writing yours.

Once the free year is active, the next question is how to study with it week to week. Our guide on how to use Gemini to study walks through a weekly system: one study notebook per exam, a diagnostic quiz to find your gaps, Guided Learning for the hard topics, and spaced quizzes you answer from memory.

Using the same plan for part-time work

Plenty of people claiming this offer also freelance, intern, or work a campus job. On Google AI Pro, Google highlights Gemini inside Gmail and Docs, which helps when you write to a supervisor or turn messy notes into a cleaner page. For the desk version, see how to use Gemini in Gmail and the Gemini tutorial for beginners.

If work and school share one personal Gmail account, keep school notebooks and work chats separate. Do not paste client files, unpublished research, or anything confidential unless your workplace allows it. The same caution applies to lecture recordings if your instructor asked you not to share them. Google’s student page also suggests career-prep uses such as improving a resume or practising interview questions from a job description. We have a fuller walkthrough on how to use AI to prepare for an interview.

When the free year ends (and how not to get surprised)

Write the end date in your calendar the day you sign up. The offer period is 12 months from the day you redeem it, and the offer itself must be redeemed by 31 December 2026.

Google’s terms say that when the offer period ends, your payment method is charged the standard monthly price for the plan in your country, unless you cancel first (or unless local law requires extra consent). In the US that listed price is $19.99 a month for Google AI Pro. Outside the US, Google lists $4.99 a month or the local equivalent for Google AI Plus. You can cancel at any time. If you cancel during the free year, Google says the membership may stay active until the end of the offer period.

For some Google AI Pro student offers and the Pro plus YouTube bundle, Google says you must reverify through SheerID, usually every 12 months, for up to 4 consecutive years. You get an email when it is time. If you do not reverify, you lose the offer and may be charged the regular membership price.

If you upgrade from the 12-month trial to a regular paid plan, Google says you cannot go back to the trial and you will not be eligible for that trial again. Signing up for a student offer while you are already on another Google One extended trial replaces the current offer. To cancel later, use your Google One or Google Account subscriptions page and follow Google’s cancellation steps before the anniversary, not after the charge lands.

Quick answers

Is Gemini free for every student?

No. The free year is for eligible students at a higher education institution who are 18 or older, who pass SheerID checks, and who live or study in a supported country. Regular free Gemini still exists for everyone else, with lower limits.

Can I use my university email?

No. Google says you must sign in with a personal Gmail account. School-issued accounts are not eligible for this offer.

Why do they ask for a card if it is free?

Google requires a qualifying payment method at sign-up. The 12 months are at no charge. After that, the same payment method is used if the plan converts to a paid subscription. That is why the cancel date matters.

I already had the 2025 student trial. Can I get this one?

The Gemini student page says students whose 2025 AI Pro trial has expired can qualify, with yearly verification. Start at the official offer page and see what your account is shown.

What if SheerID rejects my documents?

Match the name exactly, include a current date, and add a second document if your ID is incomplete. After three failed attempts you need SheerID support, not another upload on the same account. Google links to SheerID’s student FAQ from the help article above.

Does this replace paying for ChatGPT or Claude?

Not by itself. The student year gives you a lot more Gemini. It does not decide which model is better for every task. If your courses need long-document reasoning, try Claude on the same reading and compare. If your notes already live in Google Drive, Gemini has a natural home-field advantage.

Sources

• Google: College students get 12 months of Google AI free (19 August 2026)
• Gemini for Students
• Google One Help: Claim a Google One student offer
• Google One Special Student Pricing Offer Terms
• Official student offer sign-up page
• Google: AI subscription updates (Gemini Omni description)
• Google: fall 2026 AI plan updates (Gemini Spark note)

Your next step

If you think you qualify, claim the offer on the official page this week, then pick one real course and build a single study notebook from this week’s materials. That is how you find out whether the extra limits are worth keeping when the free year ends. If you want a weekly habit for using AI on actual work, not only on coursework, join Launchpad below.

You have a question that matters for work. Which supplier to shortlist, what a new rule means for your clients, how three competitors price the same service. You open a browser tab, then twelve more, and an hour later you have notes everywhere and no answer. Both ChatGPT and Gemini promise to do that reading for you. So which one should you actually use?

Here is the short answer. For ChatGPT vs Gemini for research, the honest split is this: pick ChatGPT deep research when you need control over the sources (it can stick to websites you name and downloads the report as Word, PDF or Markdown), and pick Gemini Deep Research when your work already lives in Google (it searches with Google by default, can read your Gmail and Drive, and exports straight to Google Docs). Both write a cited report in roughly 5 to 30 minutes. Neither one is right every time, so below you get a side-by-side, a 30-minute test you can run with your own question, and the checks that stop a confident wrong answer from landing in your slide deck.

Want a free starter path for using AI tools like these at work? Start on the RobotsHelp home page.

ChatGPT vs Gemini for research at a glance

When people compare the two for research, they usually mean the deep research modes, not a normal chat. A normal chat answers fast from what the model knows plus a quick search. A deep research run plans a set of searches, reads many pages, and hands you a longer report with links. Here is how the two compare, based on the official help pages from OpenAI and Google:

Features and limits change often, so treat the table as a starting point and check the help pages linked above before you decide on a paid plan.

Where ChatGPT wins for research

You decide which websites count

This is the biggest practical difference. In ChatGPT you can open Sites > Manage sites and either restrict the research to the domains you list, or prioritize them while still allowing the wider web. If you work in a field where only a few sources are trusted (a regulator, an industry body, your own company knowledge base), that control matters more than anything else. A report built on forum guesses is worse than no report.

You can steer it halfway

OpenAI lets you follow progress while the research runs and interrupt it to refine the focus or change which sources it may use. That is handy when you see it drifting into the wrong country or the wrong year. You do not have to wait 20 minutes for a report you already know is off.

The report is easy to hand on

Finished reports open with a table of contents, a list of sources used and an activity history, and you can download them as Markdown, Word or PDF. If your manager or client lives in Word, that saves a round of copy and paste and broken formatting.

Where Gemini wins for research

Google Search is the starting point

Gemini Deep Research uses Google Search as a source by default. For local questions, recent product pages and anything where you would normally just google it, that is a comfortable fit.

It can read your Gmail and Drive

Click Sources and you can add Gmail or Drive (once the Google Workspace app is connected to Gemini), upload files, or add NotebookLM notebooks. So a question like “what did we already agree with this supplier, and how do their public prices compare?” can pull from your own email and the web in one report. If you want only your own material, deselect Google Search.

The report lands where your team already works

From the Canvas panel you can export the report to Google Docs, copy it, share it, turn it into an Audio Overview to listen to on the train, or ask Gemini to build a visualization from it. On Google AI Ultra, reports can include charts and diagrams, but Google notes that this is not available when you include Gmail or Drive as a source.

It is usually quicker

Google says a report usually takes about 5 to 10 minutes. OpenAI describes 5 to 30 minutes for ChatGPT. For a quick “good enough to decide on” brief, those minutes add up over a week.

How to use ChatGPT to research a topic, step by step

A lot of people only ever type a question into a normal chat and hope. Here is a simple routine that works for most work questions. The same steps work in Gemini with its own buttons.

1. Write the decision, not the topic (2 minutes). “Electric vans” is a topic. “Should our 6-person plumbing business lease two electric vans next year, yes or no?” is a decision. The second one gets you a useful report.
2. Choose quick search or deep research. OpenAI’s own advice: use search for quick facts, and deep research when you need depth and many sources. Do not spend a deep research run on “what time zone is Singapore”.
3. Set your sources. Upload the files that matter (a price list, last year’s report). If only certain websites are trustworthy for your question, add them under Sites > Manage sites.
4. Paste a structured prompt (copy the one below and fill the brackets).
5. Read the plan before it starts. If it wanders into something you do not care about, edit it out. Two minutes here saves you a report about the wrong thing.
6. Check the sources, then decide. More on that below. It is the step that makes the report safe to use.

Here is a prompt that works in both tools:

I need to decide: [your decision question].
Context: [country/region], [who it is for], [time frame, e.g. last 2 years].

Write a report for a busy professional, max 2 pages:
1. The short answer in 3 sentences.
2. The main options, in a table: option, cost if public,
   pros, cons, source link.
3. What the most reliable sources agree on, and where they disagree.
4. Risks or reasons NOT to go ahead.
5. Three questions I should ask a real person before deciding.

Rules: link a source for every number. If you cannot find a
number, write "not found" instead of estimating. Mark anything
that is your own conclusion with [INFERENCE].

The rules at the bottom matter more than the rest. They make gaps visible instead of letting the AI paper over them with a smooth sentence.

The 30-minute test: run one real question in both

Reading comparisons, this one included, only gets you so far. The fastest way to know which tool suits you is to give both the same question from your own work.

1. Pick a question you already half know the answer to (2 minutes). That way you can spot nonsense. A good one: a supplier, tool or rule you dealt with last quarter.
2. Start both runs with the same prompt (5 minutes). Paste the prompt above into ChatGPT deep research and Gemini Deep Research. Edit each plan the same way. Then let them run side by side and do something else.
3. Score the two reports (15 minutes). For each one, open five of the cited links. Does the source really say what the report claims? Count how many of the five hold up. Then ask yourself: which report would I forward to my boss with the fewest edits?
4. Decide (5 minutes). Higher source score wins. If it is a tie, pick the one whose output fits your workflow: Word and PDF downloads, or Google Docs.

Say you run a small marketing agency and ask both, “Which three email platforms suit a 10-person agency on a tight budget, and what do they cost per month?” The useful report is not the longest one. It is the one where the prices match the vendor pages when you click through, and where it says “not found” instead of guessing. Run that test once and you will know more than any ranking can tell you.

What both get wrong, and how to protect yourself

OpenAI is open about the limits. In its launch post it says deep research “can sometimes hallucinate facts in responses or make incorrect inferences,” may struggle to tell authoritative information from rumors, and often fails to convey uncertainty accurately (OpenAI, Introducing deep research). The same kind of problem applies to any AI research tool, Gemini included. A polished report with footnotes still needs your eyes.

• Open the link behind every number you plan to repeat. Prices, percentages and dates are where errors hide.
• Check the date of each source. A 2022 price page can look current in a report written today.
• Watch for one source doing all the work. If five claims all point to the same blog, you have one opinion, not five facts.
• Be careful with client data. Deep research in ChatGPT follows the same data controls as normal chats, so check your settings before you upload client files. With Gemini, adding Gmail or Drive means it reads your own mail and documents, so only connect what you are comfortable with. On a work account, ask IT what is allowed.
• Remember where reports are saved. In Gemini you can only find past reports if your Keep Activity setting is on. In ChatGPT, deleting the chat also deletes the report.

Free or paid: what you get for research

Both tools tie deep research to plan limits rather than a simple yes or no. OpenAI says usage varies by plan, availability depends on your plan and country, and an in-product counter shows how many tasks you have left. Google says there are daily limits and that Google AI Pro and Ultra users get higher limits, and you need to be 18 or over and signed in. Start on what you have. If you run out of research reports every week, that is a real sign you would use a paid plan. If you want the full price comparison, our guide to ChatGPT vs Gemini Pro walks through what the paid plans add.

So which one should you research with?

• Choose ChatGPT if you need to limit research to trusted websites, want to steer the run while it works, or hand reports on as Word or PDF.
• Choose Gemini if your work lives in Gmail, Drive and Google Docs, you want Google Search as the base, or you like listening to a report instead of reading it.
• Use both for big decisions. Run the same prompt in each and compare where they disagree. The disagreements are exactly the points you need to check yourself.

If research is part of a bigger job, these guides go one step further: how to use AI for market research turns a report into a one-page decision brief, ChatGPT vs Claude vs Perplexity covers the search-first option, and ChatGPT vs Claude vs Gemini compares the three for everyday work beyond research.

FAQ

Is ChatGPT or Gemini better for research?

Neither wins every time. ChatGPT gives you more control over which websites it uses and exports to Word or PDF. Gemini starts from Google Search, can read your Gmail and Drive, and exports to Google Docs. Run the 30-minute test above with your own question to see which fits your work.

Can I use deep research for free?

Both companies tie deep research to plan limits that change over time. Check the usage counter in ChatGPT and the remaining reports notice in Gemini. Paid plans give you more runs.

How long does a deep research report take?

OpenAI says 5 to 30 minutes for ChatGPT. Google says usually about 5 to 10 minutes for Gemini, and longer for complex reports. In both you can leave and come back when it is done.

Can I trust the citations?

Treat them as a shortcut to the source, not as proof. Open the link, find the sentence, and check the date. OpenAI itself warns that deep research can make incorrect inferences and struggle to separate solid sources from rumors.

Should I use deep research for quick questions?

No. For a quick fact, a normal chat with search is faster. Save deep research for questions where you would otherwise open ten tabs and read for an hour.

Sources

• OpenAI Help Center: Deep research in ChatGPT
• OpenAI: Introducing deep research
• Gemini Apps Help: Use Deep Research in Gemini Apps

Pick one question that has been sitting on your list this week, paste it into the prompt above, and run it in both tools before lunch. By the afternoon you will know which one you trust. For more workflows like this, from research to small automations, join the free Launchpad.

You signed up for the certification. The exam is in six weeks, the study guide is 400 pages, and you only have the hour after the kids are in bed. You already pay for one AI tool, or you are about to. So which one actually helps you learn: Claude or Gemini?

Here is the short answer. For studying, Gemini is the better pick when you want ready-made study tools: quizzes, flashcards and practice tests you can generate from your own notes, a Guided Learning mode that teaches step by step, and Gemini Notebook for working through a stack of readings. Claude is the better pick when you study by reading hard material and arguing with it: it takes long PDFs, it pushes back on fuzzy answers, and a Project keeps your course files and tutor rules in one place. If your budget allows only one, choose by how you learn, not by which model topped a chart this week. This page shows you how to decide in 30 minutes with your own material.

Want short, practical lessons like this every week? Start free on the RobotsHelp home page.

Claude vs Gemini for studying at a glance

Both tools change their menus often. This table reflects the makers’ own help pages as of November 2026. Check the linked pages if something on your screen looks different.

Sources: Claude Help: upload files, Claude Help: use skills, Gemini Help: quizzes, flashcards and practice tests, Gemini Help: limits.

First, decide what kind of studying you are doing

“Studying” covers two very different jobs, and the tools split along the same line.

• Remembering a lot of facts. Certification terms, regulations, product specs, vocabulary for a new language. Here the work is repetition and self-testing. You want a tool that turns a chapter into 30 flashcards and a quiz without fuss.
• Understanding something hard. A pricing model, a statistics method, a legal argument, a new codebase. Here the work is explaining, getting corrected, and explaining again. You want a tool that will not let you get away with a vague answer.

Most real study weeks need both. But one of them usually dominates. Name it before you pay for anything.

Where Gemini wins for studying

Study tools that are already built

In the Gemini app you can type “Create flashcards about…” or “Create a quiz about…”, add your own files, and get an interactive set in the Canvas panel. Quizzes give you hints, let you ask follow-up questions, and offer more questions when you finish. Flashcards can be shuffled and read aloud. Google’s help page on study materials lists the exact prompts and controls.

That matters on a tired weeknight. You do not have to design a quiz format. You paste your notes and start answering.

Guided Learning teaches step by step

Guided Learning is a mode in the Gemini app that lays out a short plan for a topic, then walks you through it with questions, checking whether you are keeping up. Google describes it as a study partner that “won’t do it for you” in its Guided Learning announcement. It also uses images and diagrams, which helps for visual topics like process flows or network layouts.

It is also the only one of the two with published learning research behind it. In November 2026 Google reported a randomized trial with more than 300 University of Tokyo students: those with Guided Learning improved their test scores by 8.1 percentage points in a half-day study session, against 5.3 points for students using standard web tools (Google, Nov 2026). Keep in mind Google ran that study on its own product, and the group was students, not working adults.

Gemini Notebook for a stack of readings

If your course comes as ten PDFs and a slide deck, Gemini Notebook (the tool Google used to call NotebookLM) is built for exactly that. You upload the sources, then generate flashcards and quizzes from the Studio panel, choose easy, medium or hard, mark cards as “Got it” or “Missed it”, and repeat only the ones you missed (Gemini Notebook Help). Answers stay grounded in your sources, which lowers the risk of the AI teaching you something your course never said.

If you already live in Google Docs and Drive at work, this is the path of least effort. For a wider look at Gemini outside studying, see our Gemini tutorial for beginners.

Where Claude wins for studying

It reads the whole document

Claude accepts PDFs up to 1,000 pages in a chat. For PDFs of 100 pages or fewer it also reads the charts, images and graphics; above that it reads the text only (Claude Help). For a 250-page study guide, that means you can ask about chapter 3 and chapter 11 in the same conversation.

Gemini’s free plan, by contrast, has a 32,000-token context window according to Google’s limits page. Google’s own rule of thumb is that a million tokens is about 1,500 pages, so the free window holds roughly 50 pages before details start to slip. Paid Gemini plans raise that a lot. If you study long documents on a free plan, Claude is the safer bet.

It argues with you

This is where Claude earns its place. Ask it to act as a strict tutor and it will press on the part of your answer that was hand-wavy, instead of politely rephrasing it. For anything you must explain out loud later (an oral exam, a client presentation, a job interview on a new topic), that friction is the point.

Claude used to have a Learning “style” in a menu. In 2026 Anthropic retired styles and kept Learning as a skill. To use it, turn on Code execution and file creation under Settings, Capabilities, then go to Customize, Skills and add it from the directory (Claude Help: use skills). If you cannot find it, the tutor prompt further down this page does the same job on any plan.

Projects keep one course in one place

A Claude Project holds your course files plus standing instructions, so every new chat in it already knows the syllabus and your tutor rules. Project files have a 30 MB limit per file. We walk through the full setup in how to use Claude to study, including a 45-minute session you can copy.

The 30-minute test: decide with your own material

Reviews and benchmarks will not tell you which one suits your brain. Your own study material will. Do this once, tonight if you can.

1. Pick one chapter you have to learn anyway. Ten to twenty pages is enough.
2. Minutes 0 to 10, Gemini. Upload the chapter, then ask for a 10-question quiz. Answer every question without looking.
3. Minutes 10 to 20, Claude. Upload the same chapter. Paste the tutor prompt below and let it question you on the same topic.
4. Minutes 20 to 30, close both. On a blank page, write down everything you remember about the chapter. Then check it against the source.

Ask yourself two questions. Which session made you think harder? Which one would you actually open again tomorrow night? The second question matters more. The best study tool is the one you keep using.

This blank-page step is not a gimmick. Testing yourself beats rereading for long-term memory, which is one of the most replicated findings in learning research (Roediger and Karpicke, 2006).

Prompts that work in both

Copy these, swap the bracketed parts, and use them in whichever tool you pick.

The tutor prompt (use it in a Claude Project’s instructions, or at the start of a Gemini chat):

I am studying [topic] for [exam or goal]. Act as my tutor.
Do not give me the answer first. Ask me one question at a time.
If I am stuck, give a small hint, not the solution.
When my reasoning is wrong or vague, tell me exactly which part and why.
After five questions, ask me to explain the main idea back in my own words.

The quiz prompt (works well with an uploaded chapter):

Using only the file I uploaded, create a 10-question quiz.
Mix multiple choice and short answer. Make three questions about
the ideas people most often confuse. Do not show answers until I reply.

The gap finder (after a quiz):

Here are my quiz answers and the ones I got wrong: [paste].
List the 3 concepts I am weakest on, in order.
For each, give me one plain-language explanation and one new question to test it.

A real example: studying for a project management exam

Picture Sanne, an operations lead preparing for a project management certification after work. Her material is a thick handbook plus her own notes from an online course.

• She puts the handbook chapters and her notes into Gemini Notebook, and on weeknights runs 20 flashcards from the Studio panel, repeating only the ones she missed. That handles the terms and process names.
• On Sunday she opens a Claude Project with the same handbook and the tutor prompt, and talks through one scenario question at a time: why this risk response and not that one. That handles the judgement questions, where memorizing a term does not help.

She did not need to pick a winner. She needed each tool for the job it does best. If she could only pay for one, she would pick based on which half of the exam scares her more.

What both get wrong (and how to protect yourself)

• They can be confidently wrong. Both can produce a quiz answer that does not match your course. For exams with an official body of knowledge, check any surprising answer against your source. Gemini Notebook and a Claude Project with your files reduce this risk; a general chat with no files raises it.
• They make it easy to skip the hard part. If you ask for summaries and read them, you will feel productive and remember little. Make the tool ask you questions instead.
• Exam rules apply. Never use either tool during a proctored exam or on graded work unless the rules allow it. Use them to prepare.
• Watch what you upload. Course material is usually fine. Your employer’s confidential documents are another matter. Check your company’s AI policy before uploading internal files, and see how each plan handles your data in our Claude vs Gemini Pro comparison.

Free or paid: what you get for studying

You can test both for free. Gemini without a paid plan includes Guided Learning, quizzes and flashcards, within standard usage limits, and Gemini Notebook is available on personal Google accounts. Claude’s free plan lets you upload files and chat with a tutor prompt, with tighter usage limits than the paid plans.

When you are ready to pay, both main plans cost about $20 a month in the US: Claude Pro on Anthropic’s pricing page and Google AI Pro on Google’s Gemini subscriptions page. Google also sells a cheaper AI Plus plan with a larger context window than the free tier. Prices differ by country, so check the page for yours.

A practical rule: if you hit a limit during your 30-minute test, that tells you which free plan will frustrate you over six weeks of study.

So which one should you study with?

• Pick Gemini if your exam is mostly facts and terms, you like visuals, you already work in Google Docs and Drive, or you want quizzes and flashcards without building them yourself.
• Pick Claude if you study from long, dense documents, you need to explain and defend ideas, or you learn best when someone pokes holes in your reasoning.
• Use both if you can: Gemini Notebook for drilling, Claude for understanding. You can study this way on the free plans alone.

If you want a broader comparison for everyday work tasks, not just studying, read ChatGPT vs Claude vs Gemini. And if your study material is mostly PDFs, our guide to turning a PDF into notes with Claude shows the reading side step by step.

If you pick Gemini and you qualify as a student, you may not need to pay the usual Google AI Pro price. See how to claim Gemini for students free for a year, including eligibility and the billing catch at the end.

Picked Gemini? The step-by-step guide on how to use Gemini to study shows how to set up a study notebook, turn each chapter into quizzes and flashcards, and space your reviews across the week.

FAQ

Is Claude or Gemini better for studying?

Neither wins for everyone. Gemini has more ready-made study tools (quizzes, flashcards, practice tests, Guided Learning). Claude handles longer documents on its free plan and is stronger as a demanding tutor. Pick by whether you mainly need to remember or to understand.

Can I use Claude or Gemini to study for free?

Yes. Both have free plans that are enough to run the 30-minute test above. Gemini’s free plan includes Guided Learning, quizzes and flashcards. Claude’s free plan has lower usage limits, so long sessions may pause until your limit resets.

What happened to Claude’s Learning mode?

Anthropic moved it from the old styles menu to a skill in 2026. Turn on code execution and file creation in Settings, then add the Learning skill under Customize, Skills. The tutor prompt on this page gives you the same behaviour without any setting.

Is Gemini Notebook the same as NotebookLM?

Yes. Google now calls it Gemini Notebook. It still works the same way: you upload sources and it answers, quizzes and makes flashcards based on those sources.

Should I let AI write my study notes?

Let it draft a structure, then fill it in yourself. Notes you write in your own words stick better than notes you only read. Use the AI to test you on them afterwards.

Pick one chapter, run the 30-minute test tonight, and you will know which tool earns a place in your study routine. Want a short, practical AI lesson for work in your inbox each week?