๐จ NO MORE SPENDING HOURS CREATING DOCUMENTS. NotebookLM can now...

Step 1: Open NotebookLM.
Step 2: Upload your sources (PDFs, raw numbers, YouTube links) or let it find research gaps.
Step 3: State your format and goal (e.g., "build an investor pitch deck" or "export an Excel file").
Step 4: Download the live, editable file directly to your computer.
Upload raw numbers or operational logs. The AI runs code in a cloud sandbox to clean formatting, do exact math, and build live spreadsheets.
What it does:
โณ Runs Python code in a private cloud computer to calculate figures without hallucinated math.
โณ Automatically cleans and reconciles mismatched formatting across different data sources.
โณ Generates live Excel (.xlsx) files complete with editable cells and working formulas.
โณ Exports structured datasets directly into CSV files for external tools.
Starter prompt: "Take these uploaded ad spend and newsletter signup logs, reconcile the dates, calculate channel conversion rates, and export an Excel spreadsheet."
Turn messy folders of research, notes, or product plans into structured presentation decks.
What it does:
โณ Converts long document libraries into complete, logically ordered slide decks.
โณ Formats slides with clear structures (Title, Problem, Financials, Call to Action).
โณ Outputs real, fully editable PowerPoint (.pptx) files instead of static image previews.
โณ Tailors slide tone and content based on your specified target audience.
Starter prompt: "Build me an investor pitch deck from all the launch research in this notebook and output it as a PowerPoint file."
Have the AI analyze your data, build visual charts, and write up a formatted executive report.
What it does:
โณ Synthesizes findings across hundreds of source pages using Gemini 3.5 reasoning.
โณ Generates and embeds custom visual data charts straight into the report.
โณ Compiles multi-source findings into polished, layout-formatted PDF documents.
โณ Supports direct export to Word (.docx) for quick editing and sharing.
Starter prompt: "Analyze the data in this notebook and compile a 1-page PDF executive report complete with a line graph showing our weekly open rates."
Stop hunting down sources manually. Let the AI inspect your notebook, identify missing information, and retrieve targeted web sources.
What it does:
โณ Reads your existing notebook and flags specific missing data (e.g., market size, valuation metrics, or operating costs).
โณ Uses agentic web search to find targeted external sources that fill those exact gaps.
โณ Searches specifically for opposing viewpoints to ensure balanced, unbiased research.
โณ Protects source quality by requiring your manual approval before importing anything.
Starter prompt: "Based on what is already in this notebook, what critical data sources or opposing arguments are we missing? Find them for me."
โณ Define the persona: Don't just ask for "a deck"โask for an "investor pitch deck tailored for a pre-seed raise".
โณ Keep one notebook per project: Focused notebooks keep the AI strictly grounded and prevent cross-topic clutter.
โณ Chain the workflow: Tell the chat to find missing gaps
โ analyze raw numbers
โ generate the final PowerPoint deck in one continuous thread.
The people spending hours manually reading PDFs and copy-pasting text into slides are doing it the old way.
The ones finishing in minutes are letting NotebookLM run the analysis and build the files directly.
Save this cheatsheet. Share it with someone still building research decks by hand!
