I think NotebookLM is the most underrated AI tool in the analyst's toolkit right now, and it's not even close. While everyone's arguing about which chatbot is smartest, Google has been steadily building something different: a tool that reasons over *your* documents, not the open internet. That distinction matters enormously when you're working with proprietary financials, board decks, or internal strategy docs. Here's how I've been using it: **Earnings call prep.** I upload the last three quarters of transcripts for a company, then ask NotebookLM to surface recurring themes, flag guidance changes, and identify contradictions in management commentary. What used to take me an hour takes five minutes. **Budget review synthesis.** During planning season, I load department budget narratives and have it extract the common asks, outliers, and unstated assumptions. It's like having a second analyst on the team. **Vendor evaluation.** When comparing SaaS contracts or proposals, I upload the docs and ask pointed questions about pricing structures, SLA terms, and renewal clauses. Grounded answers with citations to the exact page. The audio summary feature is a nice bonus too: I've started generating podcast-style overviews of dense reports for stakeholders who won't read the full document. I covered this more in my [[lab/NotebookLM Slides|NotebookLM slides]] experiment. If you're building an AI presentation workflow, NotebookLM belongs in the early research phase. It's where raw inputs become structured thinking. This tool deserves more attention than it's getting. ← [[Signal|Back to /signal]]