NotebookLM’s Audio Overviews generate AI-hosted podcast conversations about your documents — and Interactive Mode lets you join the conversation. Interrupt the hosts to ask for clarification, redirect their focus, or challenge their interpretation mid-broadcast. Choose between three formats: Deep Dive for comprehensive exploration, The Brief for under-two-minute executive summaries, and The Critique for harsh, honest feedback on your work. Your sources become dynamic, shareable audio content you can participate in.
Audio Overviews launched in 2024 as one-way AI-generated podcasts about your sources. You uploaded documents, clicked generate, and two AI hosts produced a surprisingly natural conversation about your material. It was impressive — but passive. You listened, they talked, and when they missed the point or spent too long on the wrong section, you had no recourse except to regenerate and hope for a better result.
In 2025–2026, Interactive Mode transformed Audio Overviews into something unprecedented: a live, interruptible conversation between two AI hosts about YOUR documents, where YOU can participate. You can ask follow-up questions mid-conversation, redirect the hosts (“focus more on the technical architecture”), challenge their interpretation (“but the data on page 12 contradicts that”), or ask them to explain a concept more simply. The AI hosts respond in real-time, adjusting their discussion based on your input. This is not a scripted podcast with a comment section — it is a three-way Socratic dialogue between you and two AI minds about material that matters to you.
NotebookLM now offers three distinct audio formats, each designed for a different purpose and audience.
Deep Dive is the original format: a 10–30 minute conversational exploration where two hosts discuss your sources in depth, making connections between documents, explaining complex concepts naturally, and surfacing insights that might take hours to find by reading alone. The tone is two knowledgeable colleagues unpacking a fascinating topic — accessible, substantive, and genuinely engaging. Deep Dives are ideal for learning, research synthesis, and building understanding of complex material.
The Brief condenses everything into under two minutes. This is not a truncated Deep Dive — it is a fundamentally different format optimized for speed and clarity. The Brief extracts the single most important insight, explains why it matters, and states what action should follow. It is ideal for stakeholder updates, quick source summaries, daily briefings, or any situation where the audience needs the bottom line without the exploration. Think of it as a morning news briefing about your documents.
The Critique is designed for feedback. Upload your draft, manuscript, proposal, or pitch deck, and the hosts provide harsh, honest, actionable critique — the kind of feedback that is hard to get from colleagues who do not want to offend you but essential for improvement before going public. The Critique identifies weak arguments, logical gaps, confusing sections, missing evidence, and structural problems. It also acknowledges what works well, but the emphasis is on what needs fixing. This is the tough-love mentor format.
Interactive Mode is not about controlling every word the hosts say — it is about dynamic direction. The best interactions follow a pattern: let the hosts establish the foundation for the first two to three minutes, then intervene strategically. Listen to how they frame the material, which sources they emphasize, and what angle they take. Then redirect.
“Can you go deeper on that point?” “How does this connect to the section on risks?” “I disagree with that interpretation — what about the data from source 3?” The hosts process your interruption and seamlessly adjust their conversation, incorporating your question or challenge into the ongoing dialogue. This creates a Socratic dialogue between you and AI about your own materials — something that was not possible before Interactive Mode.
The strategic value is substantial. Instead of passively absorbing what the AI chooses to emphasize, you become an active participant in the exploration of your own documents. A researcher can challenge an interpretation and force the hosts to reconcile contradicting evidence across sources. An executive can redirect from methodology to practical implications. A student can ask for simpler explanations with analogies. The conversation adapts to your needs in real time.
Audio Overviews are not just for personal understanding — they are a distribution channel. Every document you upload into NotebookLM can become shareable audio content, and each format serves a different audience and context.
Generate a Deep Dive about your research paper and share it with colleagues who will not read the 40-page PDF. The conversational format makes dense material accessible, and the audio format means they can listen during a commute or workout. Create a Brief from your quarterly report for the leadership team — under two minutes of essential context they can consume between back-to-back meetings. Use the Critique to pressure-test your pitch deck before the investor meeting, getting the kind of candid feedback that internal reviewers are reluctant to provide.
This transforms static documents into dynamic, shareable audio content. The research paper that sits unread in a shared folder becomes a podcast episode. The quarterly report becomes a briefing. The draft proposal becomes a reviewed document with specific improvement recommendations. Each format turns documents people should read into audio people actually consume.
Upload the documents you want covered into a NotebookLM notebook. These can be PDFs, Google Docs, web URLs, pasted text, or any supported source type. Then choose your format: Deep Dive for comprehensive exploration, Brief for executive summary, Critique for tough feedback. The format choice should match your goal — learning (Deep Dive), sharing (Brief), or improving (Critique).
Before generating, write a brief instruction note telling the hosts what to emphasize. “Focus on the methodology section and whether it’s rigorous enough.” “Explain this for a non-technical executive audience.” “Compare the findings across all three papers.” This dramatically improves output quality by giving the AI hosts a clear direction instead of leaving them to choose their own angle.
Click generate and wait. Deep Dives typically take 2–5 minutes to generate. Briefs take under 1 minute. Critiques fall somewhere in between. The AI creates a natural conversation between two hosts who have “read” all your sources. For the Critique format, expect pointed, specific feedback rather than polite generalizations — this is by design.
Once the audio starts playing, enable Interactive Mode. Listen to the first few minutes to understand the hosts’ framing and initial angle. Then start participating. Ask for clarification on a point they raised. Redirect focus to a different section or topic. Challenge an interpretation with counter-evidence from your sources. Request deeper exploration of a specific concept. Your interventions become part of the conversation — the hosts acknowledge your input and adjust their discussion accordingly.
If the first audio does not hit the right angle, regenerate with a more specific focus note. Each generation is different — the hosts will take different conversational paths, emphasize different sections, and make different connections. Use what you learned from the first listen to craft better instructions. This iterative process often produces dramatically better results on the second or third attempt.
Download the audio file for sharing. Deep Dives make excellent podcast-style content for team learning — share them in Slack channels or team knowledge bases. Briefs work as stakeholder updates that busy executives can consume in two minutes. Critiques should be shared with collaborators working on the draft so everyone has the same feedback baseline. The audio format makes complex documents accessible to people who do not have time to read them.
| Dimension | Deep Dive | The Brief | The Critique |
|---|---|---|---|
| Length | 10–30 minutes | Under 2 minutes | 5–15 minutes |
| Tone | Conversational, exploratory | Concise, executive | Direct, honest, challenging |
| Best for | Learning, research synthesis | Stakeholder updates, daily briefings | Draft feedback, pressure-testing |
| Interactive Mode | Full support | Limited (too short) | Full support |
| Audience | You + team members | Leadership, busy stakeholders | You + collaborators |
| Output feel | Two friends discussing a fascinating paper | Morning news briefing | Tough-love mentor review |
These are focus notes (instructions before generation) and Interactive Mode interventions (used during playback). Replace bracketed placeholders with your specifics.
Every prompt in this guide plus all prompts across the full category — advanced workflows, specialized use cases, and production-grade templates.
Category Bundle — one-time access
Unlock Category Prompts — $19.99ONE-TIME · 30-DAY GUARANTEE · INSTANT ACCESS
Audio Overviews are available on the free tier of NotebookLM. You do not need a paid plan to generate audio from your sources. Interactive Mode is available on the free tier with usage limits — unlimited Interactive Mode sessions require NotebookLM Plus.
Format lengths: Deep Dive produces 10–30 minutes of audio depending on source complexity and volume. The Brief generates under 2 minutes. The Critique runs 5–15 minutes depending on the length and complexity of the document being reviewed.
Generation time: Expect 1–5 minutes for audio generation, with Briefs on the faster end and Deep Dives taking longer. Audio files can be downloaded as .wav format for sharing, emailing, or posting to internal channels. Interactive Mode requires either a microphone for voice input or text input for typed interventions.
Write specific focus notes. The most common mistake is leaving the focus note blank. A blank note produces generic, surface-level audio. A specific instruction like “Focus on the methodology weaknesses and whether the sample size supports the conclusions” produces targeted, valuable content every time.
For Critiques, specify the kind of feedback you need. “Give me feedback” is vague. “Evaluate this as a skeptical Series A investor who has seen 200 pitch decks this year” gives the hosts a perspective that produces sharp, actionable criticism tailored to your actual audience.
In Interactive Mode, wait for a natural pause before interrupting. Cutting in mid-sentence produces awkward transitions. The best intervention points are when one host finishes a thought and the other is about to respond — that is where your voice fits most naturally into the dialogue.
Regenerate if the first take misses your angle. Each generation is unique — the hosts take different conversational paths every time. If the first audio emphasizes the wrong sections or takes an unhelpful angle, regenerate with a more specific focus note informed by what you heard.
Use source notes to highlight the most important sections before generating. Adding notes to specific sources in your notebook helps the AI hosts understand which parts of your documents deserve the most attention. This is especially valuable for long documents where the most important content is not at the beginning.