DEPTH · GEMINI · 04 OF 04

Building with Gemini

Deep Research, NotebookLM, Workspace, and other Google AI tools

12 min read·4 of 4 in the Gemini track

Going further with Gemini looks different than with Claude or ChatGPT, because Gemini's strengths are different. The tools on this page are mostly for gathering, reading, and synthesizing.

This page covers Deep Research (the standout), NotebookLM (technically a sibling product but worth knowing), Workspace integrations, and where to go beyond Gemini when you need it.

You don't need to try all of this. Each section answers 'what is this, and when does it actually matter?'

STEP 01 · IN THE CHAT

Deep Research: Gemini's standout

Deep Research is Gemini doing serious information-gathering on your behalf. You give it a research question. It plans a multi-step search strategy, reads dozens of sources, and writes up a structured report (sometimes 5 pages, sometimes 50) with citations throughout.

What makes it different from regular Gemini chat: it takes time (several minutes), reads more sources than you would, and structures the answer like a real briefing rather than a paragraph.

When Deep Research is the right tool

  • You're researching a topic you don't know much about and want a real overview.
  • You need to compare options across many sources (vendors, products, methodologies, places).
  • You're prepping for a meeting, presentation, or decision and want grounded background.
  • You want to understand a current event beyond the headlines.

When it's overkill

  • Quick factual questions ('what's the population of...')
  • Creative tasks (it's built for research)
  • Anything where you already know enough and just need a quick check
My take: Deep Research is the single feature most worth trying across the three chat AIs. Even on free tier (limited monthly), one well-targeted Deep Research session can save hours of your own digging.

STEP 02 · IN THE CHAT

NotebookLM: Gemini for documents you choose

NotebookLM is technically a separate product from Gemini, but it's part of the same Google AI family and it solves a problem Gemini chat handles only partially: working with a fixed collection of your own documents as the source of truth.

You upload anything (PDFs, articles, YouTube videos, audio recordings, notes) and NotebookLM only answers based on those sources. It can also generate a podcast-style audio summary of your collection: two AI hosts having a real-sounding conversation about your documents.

When NotebookLM is the right tool

  • You're studying (a course, a topic, a book) and want to interrogate a specific set of materials.
  • You're a researcher building a personal library of references that you want to query.
  • You're prepping for a complicated meeting or trip and want to absorb a stack of documents.
  • You want an audio summary of dense material you don't have time to read.

NotebookLM has a generous free tier and is one of the most underrated AI tools currently shipping. If you read or study a lot, this is worth learning.

STEP 03 · IN THE CHAT

Workspace: Gemini in the apps you already use

Gemini's deep integration with Gmail, Docs, Drive, Sheets, and Calendar is a big strength. The sidebar is one click away in every Workspace app, and it sees what you're looking at without copy-paste.

It's already in your apps. You just have to remember it's there. Most Workspace users have Gemini available and never use it. Try it for a week. It changes a lot of small tasks.

Tasks that benefit most from sidebar Gemini

  • Summarize a long email thread before reading the latest reply.
  • Draft a response to an email (Gemini sees the context automatically).
  • Pull data out of a sheet ('what was the average across rows 50-100?').
  • Generate a document outline before writing.
  • Find your own commitments inside a long shared doc.
  • Schedule something across multiple calendars without the tab-switching.

STEP 04 · BEYOND CHAT

Gemini elsewhere in Google

Gemini is also reaching into other Google products. Some quick notes on the most useful ones.

  • Search: AI Overviews at the top of Google searches. Useful for quick context; verify before acting.
  • Android: Gemini can replace the default assistant on supported phones. Useful for hands-free queries.
  • Maps and Photos: newer features keep landing. Worth poking at if you use those daily.
  • Code Assist: Gemini for developers, integrated into IDEs. Comparable to Cursor and Claude Code, with Google's own twist.

None of these are mandatory. The Workspace sidebar and Deep Research are the two most useful places to put your attention.

STEP 05 · BEYOND GEMINI

What's next: when Gemini hits its limit

  • Claude when you need writing with a real voice, or sustained nuanced thinking.
  • ChatGPT when you need image generation, voice mode, or multi-purpose breadth in one tool.
  • Perplexity when you want search-grounded answers with a different interface and citation style (close to Gemini's strengths, sometimes sharper).
  • Specialist tools when a task needs more than a chat AI: Cursor for coding, Midjourney for images, ElevenLabs for voice.
Same lesson as the other tracks: there's no winner. Use the right tool for the task. Gemini is best for live information, long documents, and Workspace tasks. Reach for it there. Reach for others elsewhere.

TAKEAWAY

Three things to try

With Gemini, going further mostly means finding the other tools (Deep Research, NotebookLM, Workspace integration) that solve different problems. Most users never find these. The ones who do get disproportionate value.

If you remember nothing else: try Deep Research once on a real question. Try NotebookLM once on a stack of documents you care about. Try the Workspace sidebar for a week. Those three sessions teach you more than ten hours of reading about Gemini.

Try the next step

Three exercises matched to Gemini's distinct tools.

Practice 1: Run a Deep Research session

(Use Deep Research mode in Gemini: usually a toggle or model selector.)

Research question: [SOMETHING YOU REALLY WANT TO KNOW MORE ABOUT: a topic, a decision, a comparison, a current event]

Specifically:
- What's the current state of [TOPIC]?
- What are the three most important things to understand?
- What are the open debates or disagreements?
- What sources should I follow if I want to keep up with this?

Take the time you need.

Practice 2: Try NotebookLM with real sources

(This is an exercise: open notebooklm.google.com.)

Create a notebook. Upload 3-5 sources on a topic you care about: articles, PDFs, YouTube videos, your own notes.

Then ask: "What are the three most important ideas across these sources? Where do they agree and where do they disagree? What's missing from this collection that would round it out?"

Bonus: ask NotebookLM to generate the audio summary. Listen to it while you walk somewhere.

Practice 3: Use the Workspace sidebar for a week

(This is a one-week assignment.)

For the next seven days, every time you open Gmail or Google Docs, click the Gemini sidebar button at least once and ask it to do something:

- Summarize this thread.
- Draft a reply.
- Find what I committed to.
- Pull this data out of the sheet.

At the end of the week, notice: did any of those tasks shift from "annoying" to "30 seconds"? That tells you whether the sidebar is worth keeping in your daily flow.

You've finished the Gemini track

Four pieces. Trust what Gemini is good at. See where it breaks. Get skillful at search-grounded prompting, long context, and the Workspace sidebar. Try Deep Research, NotebookLM, and other Google AI tools when you have a real use for them.

You've now covered all three tracks. Claude, ChatGPT, Gemini. Each one is a real tool with real strengths. Use them where they're strong; switch when they're not.

If you'd like to practice what you've learned, the guides on this site are short and hands-on. Otherwise: you have everything you need to use AI well. Go do something with it.

Next in the track →Back to Go deeperPick another tool, or browse the guides