TL;DR
You opened gemini.google.com, asked one question, and stalled. This guide gets you from that first prompt to a repeatable Workspace task in about 30 minutes — on the free plan, before you pay for anything.
- Start free. The Gemini free plan ($0) runs Gemini 3 Flash with limited access to Gemini 3.1 Pro, plus 5 Deep Research reports per month — enough to test whether the integrations help you.
- The integration is the product. Gemini lives inside Gmail, Docs, Sheets, and Drive. Judging it by the standalone chat box misses the point.
- Upgrade only after a real wall. Google AI Pro is $19.99/month (as of June 2026) and unlocks full Gemini 3.1 Pro, a 1M-token context window, and roughly 20 Deep Research reports per day.
What this covers
A clean first hour: sign in, learn Gemini’s response style, run one Gmail summary and one Docs rewrite, generate an image, then decide if the paid tier earns its keep. No fluff, no “play around.”
Key terms, current as of June 2026:
- Gemini: Google’s multimodal AI assistant and the model family behind it (Gemini 3 Flash, Gemini 3.1 Pro). Deeply tied to Google Workspace and Search.
- Google AI Pro: the $19.99/month consumer plan that was renamed in early 2026 — you may remember it as “Gemini Advanced” or “Google One AI Premium.” It unlocks Gemini 3.1 Pro, 1M-token context, expanded Deep Research, and 5 TB of storage.
- Google AI Plus: a $7.99/month middle tier (roughly 2x the free usage limits) for people who want more headroom but not the full Pro stack.
- Gemini for Workspace: the admin-enabled version inside a work account, with side panels in Gmail and Docs on Business and Enterprise plans, and a contractual no-training guarantee for your content.
Who this is for
People new to Gemini, especially anyone already living in Gmail, Docs, Drive, and Calendar. If you are a ChatGPT power user evaluating a switch, the chat patterns feel familiar but the integration story is the real difference — do not judge Gemini by its standalone chat alone. Students picking it up for coursework should start with the Gemini study workflow, which shows how to get sourced explanations and self-quizzing — one of the highest-signal first uses.
Pick the right plan first
Most beginners never need to pay. Start free, and only move up when you hit a specific wall (Deep Research caps, context length, or model quality). Here is the lineup as of June 2026:
| Plan | Price/mo (USD) | Default model | Deep Research | Storage | Best for |
|---|---|---|---|---|---|
| Free | $0 | Gemini 3 Flash (limited 3.1 Pro) | 5 reports/month | 15 GB | Testing the integrations |
| Google AI Plus | $7.99 | 3 Flash + more 3.1 Pro | More than free | 200 GB | Light daily use, more headroom |
| Google AI Pro | $19.99 | Gemini 3.1 Pro, 1M context | ~20 reports/day | 5 TB | Real daily work, big documents |
| Google AI Ultra | from $99.99 | Highest 3.1 Pro + Deep Think | ~200 reports/day | from 20 TB | Heavy/agentic users |
Sources: Google AI plan page and Google One AI plans. Prices and limits change; re-check before upgrading.
A note on Workspace accounts: in 2025 Google retired the standalone Gemini add-on, so Gemini now ships inside Business Standard, Plus, and Enterprise plans. If your work account does not show Gemini, an admin has turned it off — that is a policy choice, not a Gemini failure.
Before you start
- Choose the account you will keep. Sign in with a personal Google account (fast to set up) or a Workspace account (what you want for work, because of scope and the no-training guarantee). Switching accounts later loses your chat history.
- Name one outcome for the session. Generate, fix, review, summarize, or compare — pick exactly one verb.
- Front-load the context. Drop the relevant Drive file, Gmail thread, constraints, and a one-line success criterion into the same conversation before you start prompting.
- Block 30 minutes. Finish one real task instead of bouncing between tabs.
Step by step
- Sign in at gemini.google.com with the account you want as your long-term AI account. By default you are on Gemini 3 Flash, with occasional access to 3.1 Pro.
- Feel the response style with a throwaway chat:
Summarize the top three AI news items from the last 24 hours with one link each.Watch how it grounds answers — Gemini cites web sources inline when it pulls from Search. - Summarize a Gmail thread. Open Gmail, click the Gemini sparkle in the side panel (consumer accounts get an in-compose “Help me write”; Workspace Business and Enterprise accounts get the full side panel). Pick a recent thread and ask:
Summarize this thread in three bullets and flag any deadlines. - Rewrite in Docs. Open a Google Doc, highlight a sentence, and use “Help me write” → “Rephrase.” Compare the rewrite to your original. This is the single most common Workspace use.
- Generate one image with a specific prompt:
Generate a flat-illustration poster for a coffee meetup, mint green and cream, no text.For coding-shaped questions instead, see the Gemini coding-help workflow so you know when to stay in Gemini and when to switch to Claude.
First-run exercise
- Pick one low-risk artifact: a single Doc draft, one PDF in Drive, or one Gmail thread — not your whole inbox.
- Run the five steps above once without changing the goal, the source, or the output format midway through.
- Save the first output into a Doc called
Gemini lab notes. Mark each sectionusable,needs editing, orwrong. - Change exactly one variable on the second run — prompt wording, the attached file, or the requested format — and compare.
Quality check
- Did the output solve the original goal, or does it just look longer? “Longer and prettier” is the most common Gemini failure mode.
- Verify any number, date, link, or quote. Grounding lowers hallucination but does not remove it, especially on niche topics.
- Note the human-judgment risk you are leaving on the table: privacy of the attached doc, copyright of generated images, or tone in an outbound email.
How to reuse this workflow
- Save the prompts that worked in a
Gemini snippetsDoc. Gemini has no Custom Instructions like ChatGPT, so a saved-prompts Doc (or a Gem) is the substitute. - For recurring work (weekly status, meeting prep), build a checklist: input ready, prompt template, review pass, send.
- Keep failure samples too — especially times Gemini confidently confirmed a wrong claim. Pattern-recognize where you should not trust it.
- Re-run a small sample every 4-6 weeks. The UI moves, model versions roll (3 Flash and 3.1 Pro both shipped in 2026), and the free-versus-paid line shifts.
When the free plan is enough — and when to pay
Stay free if your tasks are short, occasional, and tolerant of Gemini 3 Flash quality, and you run fewer than 5 Deep Research reports a month. Move to Google AI Pro ($19.99/month) when you consistently hit one of these walls:
- You want Gemini 3.1 Pro quality on every prompt, not just when the free tier grants it.
- You feed it large documents and need the 1M-token context window.
- You run Deep Research more than 5 times a month (Pro raises this to roughly 20 reports per day).
If meeting prep is your highest-yield task, the Gemini meeting prep workflow chains Calendar + Drive + Gmail into a single 10-minute brief. If that saves you 20 minutes on a Monday, the upgrade math is easy. If it does not, you have a real data point.
Common mistakes
- Concluding “Gemini does not work” when an admin has simply disabled it on your Workspace account.
- Comparing free Gemini (Gemini 3 Flash) against ChatGPT Plus on advanced tasks. The fair comparison is Google AI Pro versus ChatGPT Plus, both at roughly $20/month.
- Treating Gemini as a standalone chatbot and never opening the Gmail or Docs side panel. The integration is the product.
- Letting Gemini auto-draft emails and sending them unread. The default assistant tone reads as obviously AI to recipients.
- Skipping a voice sample. Without 3-5 examples of your own writing, every reply reads like a press release.
FAQ
- Do I need to pay to start?: No. The free plan ($0) runs Gemini 3 Flash with limited 3.1 Pro access and 5 Deep Research reports a month — enough to test whether the integrations help your day. Upgrade only after one task clearly hits a quota or model-quality wall.
- What is the difference between the free plan and Google AI Pro?: Pro ($19.99/month as of June 2026) gives you full Gemini 3.1 Pro on every prompt, a 1M-token context window, roughly 20 Deep Research reports per day, and 5 TB of storage. There is also a $7.99 Google AI Plus middle tier with about 2x the free usage limits.
- Whatever happened to “Gemini Advanced”?: It was renamed. The paid consumer tier that used to be called “Gemini Advanced” or “Google One AI Premium” is now “Google AI Pro” at $19.99/month.
- Does Gemini train on my Workspace data?: On eligible Business and Enterprise plans, Google states your prompts and content are not used to train models and are not retained beyond the session. Confirm in your admin console before pasting sensitive content.
- Can my admin see my Gemini prompts?: On Workspace plans, admin audit logs can capture activity. Treat prompts like Drive content.
- Why is the answer different in Gmail versus gemini.google.com?: Different surfaces use different model defaults and prompts. Test the surface you actually use.
Related
- Gemini Workspace integration
- Gemini vs ChatGPT
- File Summarization with Gemini
- Gemini Image Generation Tutorial
- Gemini Limitations: Region, Account, and Language Gotchas to Know
- Gemini Google Workspace deep workflow
- Gemini 3 Pro Deep Research Workflow for Big Topics
Tags: #Gemini #Tutorial #Getting started