Good prompts, safe use, and your plan.

You've got the picture now: what Gemini is, and how to use it across Gmail, Meet, Docs, Slides, Sheets and NotebookLM. Two things turn that from knowledge into results: asking well, and using it safely. Then we'll wrap it into a calm plan you could start on Monday. This is the lesson that makes the rest stick.

The prompt recipe that lifts every answer

The single biggest difference between a flat answer and a great one isn't the tool, it's how you ask. A vague prompt gets a vague draft. A bit of scene-setting gets something you can almost use as is. The recipe is simple and works everywhere, the app panels, the Gemini app, NotebookLM:

• Context. Say what it's for and who'll read it. "A reminder email to clients who haven't paid their invoice."
• Role. Tell it the hat to wear. "You're a friendly but professional accounts person."
• Example. Show it your tone, or paste a sample of how you'd normally write.
• Format. Name what you want back. "Three short paragraphs, warm, with a clear call to pay."

And treat it as a conversation. The first answer is rarely the final one, that's normal. "Make it shorter", "less formal", "add a line about the due date", and it adjusts. Most of the value is in that quick back-and-forth, not the opening shot. When a prompt works well, keep it, you've just built a reusable template.

Safe use at work

The same care you'd take with any business tool applies here, and most of it comes down to two rules. First, mind the data. There's a real difference between a personal, consumer AI account and your business Google Workspace setup. In a paid Workspace business environment, your content is contractually set not to train on it and stays under your organisation's controls. So keep confidential client details, financial records and anything sensitive inside your proper Workspace tools, and out of personal accounts and random AI apps. If your business runs on Microsoft instead, the same thinking applies, and our Microsoft 365 Copilot course covers it on that side.

Second, keep a human on anything that matters. Gemini can be confidently wrong, so anything customer-facing, factual, legal or financial gets read and signed off by a person before it goes out. That one habit prevents almost every AI horror story you've heard.

This is also where Gemini meets governance. A short, clear AI policy lets your team use these tools with confidence instead of guessing: which tools are approved, what data is fine to put in and what never is, and who checks the things that carry risk. We have a whole free AI policy and safe-use course if you want to write one properly. For now, even a one-page rule pinned where people can see it is enough to start safely.

Your 90-day plan

Don't try to boil the ocean. Roll it out the same calm way that works for any change, one proven win at a time:

• Weeks 1 to 2. Start with the two easiest wins: drafting replies in Gmail and turning on meeting notes in Meet. Low risk, daily payoff, no setup to speak of.
• Weeks 3 to 6. Add Docs and Slides for drafts, and have one person trial NotebookLM on a real briefing. Share a few prompts that worked so the team isn't starting from scratch.
• Weeks 7 to 12. Write your one-page safe-use rule, build a small shared prompt pack, and bring more of the team along. Check in on what's actually saving time and do more of that.

The goal of the first 90 days isn't a transformed business. It's a few real wins, a team that trusts the approach, and a safe-use rule everyone understands. Everything else builds from there, including the deeper AI automation work that ties these tools into how you actually run.

Keep the summary handy

To make this stick, grab the free one-page summary: where Gemini lives, the prompt recipe, the safe-use rules, and a space for your first wins. Print it, stick it by the desk, and you've got the whole course in arm's reach.