Podcasters: Save hours with this surprising AI tool

The production tool you never knew you needed

Feb 10, 2025

Sometimes it feels like AI is overhyped.

But there are some tools I use which have literally saved me hours.

One new tool in particular has revolutionised the way I plan and promote my podcasts. And even though it’s been hyped heavily on social in the past few months, most people I come across are missing the most useful features.

So, in this email I’ll explain how Google’s NotebookLM is one of the most powerful tools for podcasting - or indeed any content project - you never knew you needed.

And it’s not for the reason you might think…

Know someone who needs to speed up their podcast production? Share this email with them

You’ve probably seen notebookLM on social media, though you may not have realised it.

It’s been doing the rounds for one particular feature which has got people very worked up.

That’s the ability to create an AI generated podcast, featuring two completely synthesised voices, on any topic based on sources you provide.

For example, here’s one it generated for me on the future of podcasting in the UK:

The state of the UK podcast industry - NotebookLM podcast:

1×

0:00

-14:22

File not playing? Click here to listen.

Pretty convincing right?

I’ve heard two responses to this:

“Wow I can create any podcast I want - who needs podcasters now?”
“The robots are coming, podcasting is over, run for the hills”

But both of these miss the point.

NotebookLM is not a content production tool.

It’s a learning tool.

Wait. I thought you said it saved you time in production. Now you’re saying it’s not a production tool?

That’s right. Because where NotebookLM excels is at helping you assimilate information.

Here’s how I’ve been using it to speed up my podcast processes:

Project planning

At the New Statesman we regularly produce podcasts in conjunction with external partners.

The projects invariably come with ridiculous numbers of stakeholders, and generally we’ll be working on several at any one time.

One of the hardest things about these projects is keeping on top of the complicated multi-stakeholder email chains.

Seriously, I’ve lost hours filtering through email threads to try and piece together who has done what and what actions need to happen next.

Then I had a brainwave: let NotebookLM do it for me.

I set up a series of automations to forward all my client project emails to a Google doc, which is linked to NotebookLM as a source.

And because you can have up to 50 sources in each notebook, I also add in all my briefing documents, meeting transcripts and podcast scripts.

Now any time I need to check the status of a project I just ask NotebookLM and it gives me an up to date summary.

It can also create project briefs, FAQs and timelines, as well as summarise notes, create action point lists and draft emails… honestly, for complicated projects it’s a godsend.

Content archive

This week I’ve been editing new trailers for the New Statesman podcast.

Every time I cut a trailer I think, “It’s 30 seconds long, I’ll have that edited in no time”

But it always takes FOR. EVER.

Not because the edits are complicated - but because it takes ages to find exactly the right clips from the back catalogue to make the trailer sing.

If you’re super organised you might have all your episodes transcribed already (bonus points if you’ve kept a log of trailer-friendly sound bites - but seriously who has the time?!)

Either way that’s a lot of content to listen or read through.

Why not let NotebookLM do the searching for you?

Here’s how I used it to help me make the most recent set of trailers:

I created a new notebook for each show we run. Then I uploaded all episodes for each show from the past year to NotebookLM as new sources. (It accepts audio files as sources, which it transcribes for you, or if you have the transcripts already you can add them as a text source.)

Then using the chat function I gave it the line of script I was going to use to lead in to a clip, and asked it to find me a number of quotes from the podcast that could illustrate that line.

It took a few goes for it to find a line I was happy with, and sometimes I found it more useful for inspiration than for delivering exactly what I needed first time, but overall it made the process significantly quicker than running through transcripts or listening through audio manually.

Now that I’ve got the archive in NotebookLM I will be able to draw on it whenever I have a question about topics we’ve covered in the past, or if we need to find a particular episode or clip.

Other uses of NotebookLM for podcasting:

Researching complex reports: add your source reports to NotebookLM to create a catalogue you can interrogate in the course of your research and scripting

Presenter briefing notes: add background research to NotebookLM and get it to create a bullet-point crib sheet as an aide-memoir for your presenter

Social promotion: add your latest episode to NotebookLM and ask it to give you a selection of clips you could use to promote your episode on social - or even to draft a blog post or show notes summarising the episode.

Like any generative AI tool you need to be careful to check the output to make sure it’s accurate, but I have generally found NotebookLM competent at providing summaries and first drafts for human editing.

Are you using AI and automation tools to enhance your workflow?

Or are you concerned about the rise of AI for ethical or existential reasons?

I’d love to chat about it - drop a note in the comments or reply to this email.

Access NotebookLM for free at NotebookLM.google.com.
Hit “add source” to get started - you can add PDFs, audio files, Google Docs, YouTube videos and more.
Use the chat function to ask questions about your sources.
Use the Studio section to create briefing notes, FAQs, timelines and AI generated podcasts from your sources.

Worth your time

As well as writing this newsletter every week I’ve been publishing daily updates to LinkedIn, and I’ve found it to be a fantastic way to connect with people who are interested in the same things I am - podcasting, in this case.

It’s also contributed significantly to the growth of this newsletter (hello to those of you who found me there!)

As other social media seems to be fragmenting, LinkedIn seems to be coming into its own as a platform to help grow a community around your specialist subject.

So if you’re creating educational content around professional interests, I strongly recommend exploring LinkedIn as part of your marketing mix.

I’ve listened to a LOT of podcasts and YouTube videos about how to make the most of LinkedIn but by far and away the most valuable is this one from Jay Clouse’s fabulous podcast, Creator Science.

Jay interviews Richard van der Blom, creator of the LinkedIn Algorithm Report - a 100+ page report drawing on data from 1.5 million LinkedIn posts to reveal what really works right now on the platform.

I strongly recommend listening to the episode, and also downloading the Algorithm Report. And once you’ve digested it, why not add it to NotebookLM to ask it questions or generate an AI podcast to help you retain the information?

That’s all for this week. As ever, thanks so much for reading.

If you found value in this, please do share it with a friend so they can benefit too.

Until next time,

Chris