Why Now Is the Perfect Time for AI-Powered Podcasting

In 2026, the podcast industry is undergoing an AI-driven revolution. Platforms like Spotify, Apple Podcasts, and Xiaoyuzhou are seeing continuous user growth, while the most painful part of content creation — production — is being completely transformed by tools like Google NotebookLM.

Previously, producing one podcast episode required: topic selection → scriptwriting → recording → editing → publishing, taking several days. Now, you only need a document (PDF, webpage, or even a few notes), and AI can generate a podcast that sounds like a real human conversation within minutes.

Core Tool: Google NotebookLM

What Is NotebookLM?

NotebookLM is Google’s AI research assistant. In late 2025, it introduced the Audio Overview feature — it automatically generates a two-host conversational podcast from your uploaded reference materials. Two AI hosts discuss your content in a natural dialogue format, sounding just like a professional podcast show.

Step-by-Step Guide

Step 1: Prepare Your Source Material

You need high-quality reference material. Sources include:

• Industry reports (PDF/Google Docs)
• Long-form blog posts or press releases
• Personal knowledge notes
• Research papers or whitepapers

Pro tip: The more detailed and well-structured your source material, the higher quality the generated podcast. Aim for at least 3-5 pages of text.

Step 2: Upload and Generate

1. Visit NotebookLM
2. Create a new Notebook
3. Upload your reference material (supports PDF, Google Docs, text files)
4. Click “Generate Audio Overview”
5. Wait 2-5 minutes — AI will automatically generate a two-host conversation

Step 3: Edit and Optimize

While AI-generated podcasts are already quite good, you can further polish them:

• Use Audacity (free) to trim overly long or short segments
• Add simple intro/outro music
• Export as MP3

Step 4: Publish to Platforms

Upload your finished podcast to these platforms:

Monetization Strategies: From Side Hustle to Career

1. Platform Revenue Sharing and Ad Income

Once your podcast accumulates a certain subscriber base, platforms offer ad revenue sharing:

• Spotify: Eligible for ad sharing after 5,000 valid plays
• Apple Podcasts: Earn subscription revenue via Apple Podcasts Subscribe
• Xiaoyuzhou: Brand collaboration deals, single sponsorship ranging from ¥500-¥5,000

2. B2B Content Services

This is a severely undervalued monetization path. You can offer AI podcast production services to:

• Corporate training departments: Convert internal documents into employee training podcasts
• Knowledge-sharing influencers: Help them batch-convert articles into audio content
• Legal/financial advisors: Turn professional knowledge docs into podcasts to expand influence

Pricing reference: A 20-minute AI podcast episode costs ¥200-¥800 to produce. At 8-12 episodes per month, monthly income ranges from ¥2,000-¥8,000.

3. Build Your Own Podcast IP

Choose a vertical niche (e.g., AI tool reviews, tech trend analysis), consistently produce high-quality podcast content, and build your personal brand. Once you reach 10,000+ followers, monetize through:

• Brand sponsorships (¥1,000-¥10,000 per episode)
• Paid community/course referrals
• Affiliate marketing commissions

Efficient Workflow: Automated Batch Production

Using Hermes Agent’s automation capabilities, you can achieve a fully automated podcast production pipeline:

# Pseudo-code: Automated podcast production pipeline
import os
from notebooklm_api import NotebookClient

def auto_generate_podcast(doc_path, output_dir):
    """Automatically convert documents into podcasts"""
    client = NotebookClient()
    
    # 1. Create Notebook and upload documents
    notebook = client.create_notebook(
        name=os.path.basename(doc_path),
        documents=[doc_path]
    )
    
    # 2. Generate audio
    audio_url = notebook.generate_audio_overview()
    
    # 3. Download and process audio
    audio_path = download_and_trim(audio_url, output_dir)
    
    # 4. Generate podcast metadata (title, description, tags)
    metadata = generate_metadata(notebook)
    
    return audio_path, metadata

# Batch processing
for doc in os.listdir("./source_docs"):
    audio, meta = auto_generate_podcast(f"./source_docs/{doc}", "./output")
    publish_to_platforms(audio, meta)

With a cron job, you can schedule weekly automatic scanning of new documents, podcast generation, and publishing across all platforms — truly achieving “passive income.”

Real-World Case Studies

Case Study 1: Tech Blogger’s AI Podcast Matrix

A tech blogger uses NotebookLM to compile weekly tech news summaries into documents, automatically generating 5-8 podcast episodes. Through multi-platform distribution, they earn approximately ¥3,000/month in ad revenue while driving traffic to their paid community.

Case Study 2: Law Firm’s Professional Podcast

A small law firm created a podcast series themed “AI Legal Assistant,” converting common legal questions into documents and using NotebookLM to generate podcasts. These podcasts not only boosted brand awareness but also brought 2-3 new client consultations per month.

Getting Started: Your Action Plan

1. Week 1: Register NotebookLM, upload 3-5 documents from different fields, experience the generation process
2. Week 2: Choose your most familiar domain, produce your first official podcast episode and publish it
3. Month 1: Maintain a frequency of 2-3 episodes per week, test audience response to different topics
4. Month 3: Optimize content direction based on data feedback, start exploring monetization channels

Summary

The core advantage of the AI podcast side hustle is extremely low marginal cost — once you’ve built a content production pipeline, the cost of adding each new podcast episode is nearly zero. As AI audio quality continues to improve, the competitive barrier in this space will keep getting lower, while the brand effect of early entrants will keep strengthening.

Now is the best time to enter. Open NotebookLM, upload your first document, and begin your AI podcast journey.