Google's Gemini, taught as a research assistant rather than a feature list. Seven hands-on lessons take you from your first real task on the free tier through prompt-craft, working with your own files and Gmail/Drive, web answers you can verify, generating and reading scientific visuals, and packaging recurring jobs as reusable Gems — ending with one full agentic Deep Research report you direct from plan to deliverable. Everything here runs on Gemini's genuinely capable free tier; paid plans only raise the limits.
Gemini ScientistPaste a table into Gemini, ask it to describe the trends, and have it generate a chart and a figure illustration for a slide.
A quick description plus a draft visual, using Gemini's multimodal strengths.
Paste a CSV snippet → "Describe the trend, then generate a labeled bar chart and a simple schematic figure."
Gemini Small biz AI-native SMEBuild a Gem — a saved custom assistant with its own instructions — for a task you do often, so it starts in the right role every time.
A one-click specialist instead of re-typing the same setup prompt.
Create a Gem with instructions for your task, then run it on a real example.
Gemini AI-native SME FounderFor organisations subject to GDPR or sector-specific EU data-residency obligations, use Gemini via Google Workspace (Business or Enterprise) or Vertex AI with an EU region selected — not the consumer Gemini.google.com interface.
Prompt and response data stays within the EU region and is governed by the Google Workspace data-processing agreement, satisfying most data-residency audit requirements.
In the Google Cloud console → Vertex AI → select a europe-west region before creating any Gemini API project. For Workspace, confirm your admin has enabled the EU Data Region in Admin console → Account → Data regions.
Gemini Scientist CreatorDescribe a concept and have Gemini generate a clean schematic or illustration for a slide or poster, then refine it with follow-up edits.
A draft visual in seconds you iterate on, instead of fighting a drawing tool.
"Generate a simple, labelled diagram of the cell cycle for a lecture slide — flat style, clear text."
Gemini Creator Small bizAsk Gemini to produce a structural outline or a rough first draft for a post or article, but then rewrite every sentence yourself before publishing. Treat the AI output as a research assistant's notes, not a finished product.
Your published content reflects your authentic voice and original judgement, and you avoid the reputational and platform risk of publishing AI prose that readers or algorithms flag as generic.
Prompt: "Give me a bullet-point outline for a post on [topic], including 3 angle options." Then write the post from scratch using only the outline — do not copy any generated sentences.
Gemini Founder AI-native SMEUse Gemini's Deep Research to send it off to browse many pages and come back with a structured, multi-section report you can skim and verify.
A broad first map of a topic with links, assembled while you do something else.
Pick Deep Research → "Survey the current landscape of X: key players, approaches, and open questions."
Gemini Scientist FounderEnable the "Google Search" grounding tool in Gemini Advanced so responses cite live web pages. For every cited link, open it and confirm the supporting sentence appears verbatim or paraphrase — do not assume the model quoted accurately.
Grounded responses are more likely to be traceable, but models still misquote or cherry-pick context; manual spot-checking catches these errors before they propagate.
In the Gemini web app, click the globe icon to enable Search grounding. After the response, click each citation chip and find the exact supporting sentence in the source page.
Gemini ScientistUpload multiple files and ask Gemini to compare them — pull the shared findings, the disagreements, and a combined summary using its long context.
A synthesis across a stack of papers instead of reading each one cold.
Upload 3 papers → "What do these agree on, where do they disagree, and what's the combined takeaway?"