Turn dense reports into your commute podcast.

What the AI will do

• Read the reports and relevant accessible work discussions, then prepare the sources and listening focus for NotebookLM Audio Overview.
• Import, customize, generate and check the audio where its tools allow; guide only the account or device steps you must perform. If NotebookLM cannot be used, prepare a reviewed script and synthesize it with an available audio tool.

What you can provide

• Reports or links, and any known listening purpose, language and approximate commute length.
• NotebookLM access or the minimum account actions needed to use it; fallback synthesis may also need your access.

Where you'll take part

• Confirm the report set and listening focus when unclear.
• Perform any user-only import, generation or phone-download actions, and help check playback if the AI cannot hear or inspect the recording.

What you'll get

A source-checked recording you can play on your commute, normally made with NotebookLM Audio Overview. The AI states what it checked and whether offline playback was verified. If access stops production, you receive an import-ready package or checked fallback script, explicitly short of a completed podcast.

Activity context

Let the dense report become a conversation you can follow on your commute. NotebookLM Audio Overview is the first route: give it the right original sources and a clear listening focus, then check what it produced.

Start with NotebookLM Audio Overview. If it cannot be used, the fallback is a reviewed script rendered through ElevenLabs Studio or another available synthesizer. Product controls and account access may differ; the AI checks your actual setup.

For the user

Point the AI to the reports. It will find the context that matters, prepare the sources and customize the recording. It handles supported actions itself and gives you the exact material and short instructions for any account or phone steps only you can do.

Guidance for the AI

These are general instructions, not a description of this user's particular AI system, tools, access, or circumstances. Adapt the method to the user's goal and the capabilities you can actually use, while preserving the activity's essential sequence, review points, and intended result.

Start with the activity and the user's context

Read the complete activity and guide the user through it. Use the activity prompt yourself; the user does not need to paste it again. Keep the responsibilities of the user, the AI, and any colleague clear.

Begin with what the user has already supplied. Establish the intended result and resolve missing context through focused questions when needed. Work within the activity's stated scope, permissions, required confirmations, and stopping points, including any deliberate order for introducing information. Make progress with the available information and reasonable, clearly identified assumptions where the activity allows them.

Check the capabilities needed for this work

Inspect relevant available tools, connectors, and data. Use small, relevant, non-destructive checks to establish whether necessary access or a capability actually works. Naming a product does not establish access to it. Let this activity determine what needs checking; do not impose a universal preliminary questionnaire or a full system audit.

When useful access is missing:

• Explain what you cannot reach and how that material would help.
• Investigate practical setup, connection, export, or other access alternatives before asking the user to gather material or troubleshoot.
• If progress depends on access, request only the specific user action or permission needed to enable it or provide the material another way. Continue useful independent work while awaiting the response.
• Offer to walk the user through an export or upload. Welcome files, links, copied text, rough notes, or an interview as supported by your system and the activity, and use the best available route. If the missing access is optional, continue with what is available.

Ask focused questions and respect the answers

Ask questions only when necessary to complete the work well or satisfy the activity's required review points. Build on answers already supplied. Ask one question at a time, or a small group of closely related questions. Wait on work that depends on the answer, and continue useful work that does not. Explain unfamiliar language with examples where helpful.

Confirm answers or decisions you supply on the user's behalf before treating them as the user's final answers or decisions. This does not require separate approval of every sourced finding; follow the activity's specific review requirements.

Keep evidence and independent contributions distinct

Distinguish evidence, inference, and missing information. Check your work against the source material.

Use independent agents when your system actually supports delegation and independent investigation or review would improve the work:

• Give each agent a bounded brief and the context appropriate to its role.
• Keep discovery independent until synthesis where the activity requires it. Preserve any requirement for the user or colleagues to contribute before seeing other explanations.
• Follow the activity's stated fallback and any required confirmation when independent agents or fresh contexts are unavailable. Preserve reviewer isolation where required; describe shared-context or sequential passes accurately.

Produce, check, and revise the result

Do the substantive investigation and production. Carry the activity through to the most complete, usable result within its scope, rather than ending with a plan or handing the main work back to the user. When missing information does not prevent useful work, proceed with clearly identified assumptions where the activity permits them.

Check the result against the activity's completion criteria. At each substantive delivery, state what is complete and any material unfinished work, assumptions, helpful missing information, and remaining uncertainty. Explain how these affect what the user can use, and help the user revise the result.

Preserve the activity's stopping point. Do not turn a rehearsal, investigation, or test-design activity into execution that the user has not authorized.

Your AI prompt

Turn dense reports into a commute-ready podcast using NotebookLM Audio Overview first. Select and inspect the reports, prepare the listening focus, import the sources, generate the recording, check it and make it playable. Use an authored-script and speech-synthesis fallback only when NotebookLM cannot be used. Ask necessary questions one at a time and wait on consequential unresolved choices.

1. Find the right reports and listening purpose

Start with supplied reports, project names and questions. Search accessible documents, related emails and chats, and meeting transcripts for report versions, questions people are trying to answer and why these findings matter to my work. Inspect the reports themselves, including consequential tables and qualifications. Use surrounding discussion to choose emphasis; distinguish that context from report findings.

Keep a source list with titles, dates, versions and the pages supporting the main points. Try available access before asking me to collect material. If a report is unreadable or missing, investigate a usable export or upload and request only what is necessary. Do not replace original reports with an unchecked summary.

Establish my listening purpose, subject familiarity, language and approximate time available from existing context, asking where needed. Confirm consequential source or focus choices. For a long set, propose a coherent episode or series and identify coverage; do not silently omit most of the material.

2. Check NotebookLM access and prepare the listening brief

Inspect available NotebookLM or browser access and current official guidance. Documentation checked September 12, 2026 uses the title Gemini Notebook on some help pages; verify the interface I actually have. Relevant official instructions are linked in the phases below, so this workflow does not depend on another local document.

Lack of your own UI control does not make NotebookLM unusable. Prepare the exact files, source titles and customization text, then guide only the actions I must perform. Ask me to confirm results you cannot observe and resume from that state. Investigate a practical access route before moving to the fallback.

Write a listening brief containing:

• The question the episode should help me understand and why it matters.
• The intended listener and knowledge to assume or explain.
• The key findings, comparisons, figures and qualifications to preserve.
• Topics to emphasize or leave for another episode.
• Requested language, conversational style and approximate listening time.

Translate that brief into paste-ready customization text using the actual report topics. For multiple reports, use bounded source-review agents where available, with assigned documents, page references, findings, uncertainties and a coverage requirement. Reconcile the completed briefs before generation; use waves when needed. Without agents, use bounded passes or fresh-session handoffs with sources, decisions and remaining coverage.

3. Import the selected original material

Create a notebook containing only the intended reports and explicitly identified context. Use Create new notebook, then Upload a source; add later material through Add sources. Google: create a notebook.

Use supported PDF uploads, Drive documents or clearly titled pasted text. Verify that the actual sources appear and inspect important passages after import. Drive imports require viewing access; Google-file comments and footnotes are not imported. Where these matter, supply their necessary content separately and label its origin. Check freshness or synchronization before relying on an imported version. Google: source imports.

Keep source selection deliberate. Desktop chat checkboxes do not establish what the audio generator will use. Inspect the audio controls; a dedicated notebook with only the intended sources avoids that ambiguity. Mobile audio customization supports selecting sources. Google: mobile controls.

4. Customize and generate the Audio Overview

Open Studio → Audio Overview. Choose a suitable format: Deep Dive for a two-host discussion, The Brief for short takeaways, or Critique or Debate when that serves the listening purpose. Set the language and use Shorter, Default or Longer where offered; documented length controls are English-only. Paste the prepared focus instructions and generate. Inspect the completed result before claiming generation succeeded. Length preferences do not guarantee a precise commute duration. Google: Audio Overview.

If a user-only click or login is required, provide the exact next action and wait for its result. Preserve the source package and settings across that handoff. If NotebookLM truly cannot be used, explain the actual blocker and proceed to the fallback below with my agreement where it changes the tools or permitted use of the material.

5. Check what the recording says and make it playable

Listen to or otherwise inspect the actual generated audio using available capabilities. Check major claims, numbers, attribution and qualifications against the original reports, and detect missing central conclusions or audible glitches. A transcript can support content checking but cannot establish sound quality. Use notebook chat to locate supporting passages and open citations in context; treat its answer as another claim to verify. Google: chat citations.

Give a fresh reviewer the recording or transcript and source passages without your production rationale. Ask for specific mismatches and omissions, then adjust sources or customization and regenerate when needed. Disclose shared-context review or audio you could not inspect. Request only the listening checks that require me; do not label unchecked audio source-checked.

Download the web recording or use the mobile app's Download control for offline listening. Mobile downloads remain inside the app under Downloaded; background listening is supported. Check the beginning and actual duration, then verify offline playback on my intended device or ask me to do that brief check. Distinguish a file download from confirmed phone availability. Web download, mobile listening.

6. Fallback only when NotebookLM cannot be used

Write a complete spoken script based on the reviewed source set: an opening question, clear explanations, relevant examples, transitions and a useful ending. Preserve uncertainty and attribute important claims. Keep a separate source map, and have a fresh reviewer check the actual script against the reports. Repair it before synthesis.

Use ElevenLabs Studio or another available synthesizer that preserves the reviewed script. In ElevenLabs, use Upload for the script file or New blank project for an audio project, inspect the text, choose a narration voice, generate and play it, fix flawed passages and export MP3 or WAV. Do not use Generate a Podcast: its GenFM route writes a new script. ElevenLabs Studio.

Check exported audio and commute playback as above. Deliver the recording, sources and actual review status. If production stops at an import-ready package or a reviewed script, deliver that useful intermediate result and name the exact remaining action. It is not a completed podcast until playable audio exists.

Done when

You have playable audio, normally a NotebookLM Audio Overview, with its source checks and download or offline-playback status stated. An import-ready package or script without generated audio is a clearly labeled intermediate result.

These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.