In this episode of In-Ear Insights, the Trust Insights podcast, Katie and Chris discuss a shifting content landscape where algorithms act as your primary readers. You’ll discover why machines consume your content while humans scroll feeds. You’ll learn how to transform any written idea into audio, video, and short clips during your regular workflow. You’ll see which free tools handle the heavy lifting while you focus on your core message. You’ll follow a simple roadmap that aligns your content with the group that drives real results.
00:00 – Introduction
02:15 – The hidden audience driving your traffic
06:30 – Why repurposing feels like double work
11:45 – The purpose first approach to content
16:20 – Free tools that automate your workflow
22:10 – Building clips without cloud costs
26:45 – Choosing the right software for your budget
31:00 – When to skip the overhaul entirely
35:30 – Call to action
Watch the full episode to start building a content workflow that saves time and reaches the right audience.
Can’t see anything? Watch it on YouTube here.
Listen to the audio here:
- Need help with your company’s data and analytics? Let us know!
- Join our free Slack group for marketers interested in analytics!
[podcastsponsor]
Machine-Generated Transcript
What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.
Christopher S. Penn: In this week’s In Ear Insights, let’s talk about content remixing and optimization. Now I know, exactly. But it is relevant to today’s topic, which is how do we take any content and turn it into any other content? So, Katie, why don’t you lead us off with where things are today?
Katie Robbert: Content repurposing. Again, new tech, old problems. This is not something new that we’re trying to solve for marketers, content creators, and humans in general have always been trying to find ways to repurpose content. So if you think even like early on with early storytelling, stories were told and augmented through generations to bring about new information to share with a new audience. That is the essence of content creation: storytelling to convey information, whether it’s entertaining or educational, whatever it is.
And so we’re always looking for new ways to repurpose things that have resonated with people, things that have landed, to recreate the things that worked. And so this started first with the transmedia content framework, which essentially, in a nutshell, just means how can you use one piece of content across a bunch of different media platforms? And so you have video, blog, audio, et cetera. And so now with AI, it’s not that the problem has changed, it’s that the solutions to repurposing content have evolved. And so we as marketers and content creators need to figure out where these new AI tools fit into our transmedia content repurposing framework or our content transformation.
Christopher S. Penn: And part of the challenge of this is that we’re making content now for two different audiences. We’re making it for the humans who consume the content, but we’re also making it for the machines to train them on our information. Because as we’ve talked about in past episodes of both the podcast and the live stream, machines are like three quarters of our audience now. There’s more machines than humans consuming our content.
So if we follow that old maxim of “make content your audience actually wants,” if our audience is 3/4 machines, we have to put some effort into creating for them. That’s where the transmedia or the content remixing frameworks come in handy, because if we make a piece of content, we need to put it in as many places as possible so the machines can consume it as much as possible.
Katie Robbert: So that video that you created, for example, which I certainly have some thoughts on, who is that for? Is that for the humans or is that for the machines?
Christopher S. Penn: So here’s where we’ll start. Any piece of content, anything that is useful, can be turned into any other format without substantially changing the content itself. So if you have a written blog post, maybe it’s like a thousand-word blog post about, I don’t know, the best places to vacation in Massachusetts, you could relatively, in a straightforward way, take that and using a text-to-speech model, turn it into audio which just reads it aloud. That is not even remotely new technology; that’s about 15 years old. But today’s AI models do a really good job with that.
You could then take that audio file and pair it with a video generation service of some kind to create, as you saw, the talking head. That avatar is named Finley. It is made by Google. It is part of Google Workspace. So if you are a Google Workspace customer, as Trust Insights is, you get access to Google Videos. And inside Google Videos is the AI avatar service where you can upload either text or audio and it will create a very synthetic talking head.
Katie Robbert: Okay, but that didn’t answer my question. My question was who created that? So you created that avatar?
Christopher S. Penn: Yes.
Katie Robbert: You’re talking about two different audiences. You created that video. Is that video for our human audience or for a machine audience? And if it’s for a machine audience, what does that machine get from that video?
Christopher S. Penn: So that very short video would be made for a machine audience. The idea is if a longer-form piece was made into a YouTube video, that YouTube video would then get uploaded to a YouTube channel, not necessarily our main one. Maybe a secondary machine-only channel or a machine-focused channel that would contain the full transcript, the closed captions, and an extensive description to essentially add more data. And inside the YouTube channel, every AI service that’s allowed would be allowed to train on that channel’s content. So you would create this talking-head video reading aloud a blog post that would go into YouTube to be distributed as training data for machines.
Katie Robbert: I can already hear and feel the collective groans from marketers saying they barely have enough time to create content now. And I hear Chris Penn telling me now I have to create two versions of the content. I can barely get through my to-do list, and now you want me to double it? Like what the heck? Why can’t I just create content for the humans and let the machines get whatever they get?
Christopher S. Penn: Good question. First, this might not double your to-do list because if you’re skilled with tools like Claude Code or Claude Coworkers, 99 percent of the workflow can be done by a machine. From the creation of the assets to the uploading of the assets. Because it’s all APIs and stuff like that. You would have to go through the work of setting up that infrastructure, but once it’s in place, then it’s just another scheduled task that the machines do on your part. The second thing, and probably the more important thing, is there was a paper back in 2024 — and it’s something that at some point I want to recreate — that demonstrated in ranking and ordering results. AI models prefer AI content over human content.
They like the higher probability stuff. They’re more likely to recommend a machine-written piece than a human-written piece. Now that granted, this is now two years old, so models have changed significantly since then. But in things like search engines and AI recommendations — you know, take your pick of all the AI visibility and SEO stuff we’ve been talking about — having machine-made content might be more recommended to humans by machines reading other machines’ content.
Katie Robbert: I’m going to need a whole map for that, because basic. Okay, let me see if I can follow this. So if I’m a human creating content, then theoretically my content is still better served for humans than machines. And that is technically what humans are saying they want: human-generated content. If I’m a human operating a machine to create content, that content is likely better suited for another machine who is then going to choose what to serve up to the human who has already said they want human-generated content, but they’re going to get machine-generated content. So really…
Christopher S. Penn: It’s machines all the way down.
Katie Robbert: Well, you know, and it’s interesting — this is a little bit of a sidebar. I’ve been seeing a lot of chatter on places like Threads where people say, whatever happened to people having blogs on their website where they just create their original content? What’s interesting is that hasn’t gone away. What’s changed is where people consume content. And so everything that we create, Chris, you and I, lives on our website, on our blog. But that’s not where people go to read the content that we’re creating. So in addition, we have to post it on different social media channels. We have our email newsletter.
It’s almost like it’s a channel strategy, which is not a new thing. AI didn’t create it. And this is the thing that sort of gets back to the point about content transformation: I’d say number one, first and foremost, I’m going to pull up something that I know people are going to be shocked by. The 5P Framework by Trust Insights: Purpose, People, Process, Platform, Performance. Before you start overhauling your content strategy and undergoing this huge content remix or transformation, first and foremost, I want you to figure out what your purpose is. Am I trying to reach a human audience or am I trying to reach a machine audience for AI visibility? Or is it a little bit of both?
And it’s okay if it’s both, but then you want to have separate requirements for each one so that you can treat each with the time and resources they need. I would start there. Your people, that’s your audience. Is it a human audience? Do I know who those people are? Do I know where they spend their time? Is it a machine audience? How do I make sure that I’m making content for the machines? What machines are they? Do I want to show up in Gemini Search? Do I want to show up in Claude Search? Do I want to just show up in like an AI overview on a web browser?
Then you have your process of creating content, which we’re now going to talk about in terms of the content transformation and so on, so forth. So I just wanted to acknowledge that because speaking for the average marketer, this feels like a lot, and it’s not a new topic. In terms of having to have machine-friendly content, but now even more so we’re seeing that. I mean, we just published a paper on this last week, Chris, about how people are using generative AI and what the potential is. And so what we’re seeing in a nutshell, and we talked about this on last week’s podcast — which you can get at the Trust Insights AI podcast and our livestream, the Trust Insights AI YouTube channel — we talked about the potential for AI usage in jobs.
And what we learned with the original research that Claude published was that it was just looking at chat and what quote-unquote questions people were asking, which then led to the conclusions of what jobs will AI be able to take over? It’s all the research is in the paper, but my point being is that people are turning to these large language model chats to do their web search whether they realize it or not. Because a lot of companies like Google are making it the default. You’re searching with Gemini, you’re not searching with the Google web browser anymore. And so users are being forced into using these large language models whether they realize it or not. So marketers like us need to figure out: okay, if that’s where people are consuming content, how do I get my content there? That wasn’t in my plan.
Christopher S. Penn: One of the first things, and I’m glad you brought this up, is that one of the first things you should do to the 5P Framework by Trust Insights is ask: do you know who your audience is? That is a genuine real question. And go back to our live stream on our YouTube channel. This is for example the AI metrics for the last seven days for the Trust Insights website. We have 105,000 total requests, 82,000 allowed, and just a high level. Google sent 29,000 requests, of which 1,500 were actual real people. Microsoft sent 6,600 through Copilot, of which 141 human beings got through.
Our audience is machines by a landslide on our website. And that, Katie, goes back to your point about saying if you’ve got a blog, who’s the audience reading it? It is machines. If you have a blog, your audience is the machines, hands down, no question. And so when it comes to content repurposing then if you are putting content out, and this again goes really well with your point, it depends on where you’re publishing. If you are publishing on your website, your blog, this is your audience. You are making content for machines. Which means that you should have all the things that machines are expecting. Good schema, good markup, good JSON-LD, good inverted pyramid content structure, all that stuff. Because here’s who’s coming to read. Now if you are in places like TikTok or whatever, you’re still catering to machines because there’s an AI algorithm in a way, but the human audience is at least a step closer than there. So your first step should be to look at your data and see how much of your audience is a machine.
Katie Robbert: What a novel idea. It’s almost like you want people to be data-driven, but again I know I keep reiterating that new tech doesn’t solve old problems like needing to look at your data to figure out where your audience is and what they want. That should be table stakes. Like you should be doing that. That’s your foundation. Looking at the information that you have in order to figure out what you need to do next is 100 percent where you should start. And then because that’s actually probably going to help you define an even clearer purpose. Because if you don’t, and if you’re just guessing at where your audience is, the purpose is really kind of irrelevant because it’s made up, it’s hypothetical. If you’re going to spend your time overhauling your content transformation, you want to make sure you’re doing the right things.
And so for us, we know based on our data that our website is for the machines. And our data has been indicating that for a while in terms of traffic to our website, engagement rates, bounce rates, those kinds of things. That’s very clear that we’re getting more machine traffic than we’re getting human traffic. That said, our email newsletter is where the humans are. And so we can create content specifically for the newsletter first, like that can be the priority. And then whatever’s left or whatever other resources we have can go onto the website, onto social media. But if our goal, if our purpose is that human audience, then we need to focus on our newsletter.
Christopher S. Penn: Exactly. And at the same time, use content repurposing frameworks to create machine-optimized stuff. And here’s the thing, and this is something that is on our to-do list for either late this summer or during the holiday season. Our website architecture itself, the technical pieces, needs an overhaul because I just did a really large overhaul of my own website early over the holidays. I mean, I completely redid everything on it, just hammered together all new pieces and things, and the changes: I have almost quadrupled the amount of traffic that I get to my website, mostly from AI. Because I had a stupid theme and I’m like, okay, fine, Claude, you’re going to build me a new one. It still isn’t great looking because I’m not a designer, but in terms of the traffic it attracts from machines, it works much better. And so our remit, now that I’ve proven that it works out by breaking down my own website and their being a proven process, it’s now time to do it for the Trust Insights website so that we get more machines in tools like Copilot, which contain our audience. Because Microsoft Copilot serves the enterprise audience and that’s where our customers are.
Katie Robbert: Side note: I definitely have a process for creating a branding guide using Claude. If you ever need one for your own personal website, or if you’re out there listening and would like one, definitely give us a shout out. All right, so let’s talk about the actual effort of the content transformation. And so as we started with this episode, we have a framework on our website called the Video First Transmedia Framework. And so the idea at the time was that you could take any given video and transform that video into a bunch of different types of content for different reasons. That still holds. What’s changed are the tools. And theoretically, doing this effort to transform your content should be easier.
So, for example, Chris and I record the podcast on StreamYard, which is our video service, and then we post it to our YouTube channel, which you can get at TrustInsights.ai YouTube. That’s one version of the content. We also, from that podcast recording, take the audio file that goes up on the different podcast providers, you know, so your Spotify, your iHeartMedia, your Apple Podcast, whatever. Then we also have the transcript, which turns into text, cleaned up text, which becomes the blog post, which goes up on our website. Then we also have shorter versions of the podcast, the reels and the shorts, which go on different social media channels. So it’s just like one example of how we use one video, one recording, in multiple different ways. That’s the idea. The thing that Chris is probably going to school me on, because this is not something I pay close attention to, is what are the tools available to make this process even easier? Because recording a video once, like that’s the easy part. Repurposing the content correctly and thoughtfully, that’s the hard part.
Christopher S. Penn: It is. And we also, as a company and as humans who have espoused our values about how we want to do stuff — and one of those values is being responsible in the use of AI — a big part of that process just happens on my laptop because Katie was nice enough to let me buy the nicest laptop available at the time. And so once that video appears, it then goes through transcription, and we get what’s called an SRT file, a captions file. That captions file is then fed through a piece of custom code that we wrote in Python with Claude Code that does three things. One, it finds the smartest thing Katie said in an episode — specifically Katie, between 30 and 45 seconds long, or 30 and 59 seconds long. It identifies that and then it calls a command on my laptop called FFmpeg, which is a local piece of software that takes the original video file and snips out the relevant part where Katie is talking. If you’ve noticed in all the reels from this podcast ever, it has always been Katie’s stuff first.
I almost never have a place in the reels that then gets put into for now Adobe Pod, Adobe Media Encoder to turn into a 9 by 16. That’s going to be changing to just a local tool as well. And the SRT file is then taken by the local AI model. We use Quinn 3.6 to turn it into the cleaned transcript. So all of that happens locally in about 40 minutes on my laptop. I just drop in the file and then walk away and do something else. And 40 minutes later it’s ready. Then to your point, all those pieces get uploaded. But there’s a local AI model that runs on a Mac called with a server called OMLX that does all the computations to figure out what are all the pieces that need to be made. I don’t do that because that’s not a good use of my time. And as of like two versions ago, which is now about six months ago, what came out of AI was good enough and it wasn’t creating net new content, it was just atomizing. Now today’s AI tools are so capable that you do not have to start with video. And that’s what changes in the AI. First, content remixing is if you’ve got an idea and a piece of content in any form, you can now remix and extend it. You could say here is an infographic I made. Machine, turn it to a blog post. It will read through and it will draft it. And then you’ll have to clean up the AI slop. And then Machine, read aloud that blog post. Now you have an audio file which becomes a podcast. Machine, turn this audio file into a video. Maybe it’s a headline-style video, it’s just an audiogram. Or maybe it’s a synthetic talking head like Finley that reads it aloud. Machine, from there turn it into reels and shorts and so on and so forth. And the thing about that is a lot of folks who are only casual users of AI don’t know that if you see a piece of software or a service that you think does something useful, you can literally — as Katie wrote in the Inbox Insights newsletter — you can literally just ask Claude: hey, how do I make this for myself on my computer? I have a version of a clipping service now that I’ve been tinkering with for six months that it finally is functional. It can take any video and just turn it straight into those tech-talk style clips without using cloud AI, which means I don’t use a data center, I don’t cost massive amounts of electricity, it’s all renewable because it’s on my house. The only fresh water usage is the water I drink while I’m watching at work.
Katie Robbert: For those who aren’t looking to necessarily build their own, can you quickly run through some of the better options that are available for repurposing content or generating video? So one of the things that we’ve been talking about with our clients, the challenge that a lot of our clients have, and a lot of people have in general, is just time. And so they have a lot of content like written content, blog posts, they might have videos or B-roll, whatever it is sitting around and they want to do more with it. And we’ve been sharing with them ways to do this that aren’t super high tech. Can you talk through some of those lower tech options that are available to people who aren’t necessarily looking to build a local model or custom code something?
Christopher S. Penn: So some tools that do this really well out of the box. Google’s AI Studio has amazing text-to-speech. You can just paste a blog post in there and hit speak, and just grab the recording and it’s done. It’s in a web browser, it uses Google’s cloud, and you get a small amount of free usage. And by the way, your data is consumed. Not that it really matters for blog posts, but your data is consumed and used for training, which is not necessarily a bad thing. That’s an excellent choice. See, it’s silly, but you can if you wanted to get like a super low-tech and cheap option. You could literally paste a blog post into a tool like ChatGPT or Claude or whatever on your phone and say read this back to me and hit screen record. You would get the audio file out of it. That is super, super fast. And almost all computers have a text-to-speech option built into the operating system. Apple has it right in the OS. Windows has it right in the OS. So you could literally highlight text and say read it aloud. Now you have to wait and record it. But there’s a way to do that in terms of video editing if you don’t have a huge budget. The piece of software that I would recommend is DaVinci Resolve. This is a tool that Hollywood uses, and they make their money mostly from selling hardware like the $25,000 video editing station. The software is free and is very good. There’s a pro version that uses their hardware that almost nobody needs. It does have a learning curve. But you can take a screenshot as you’re working and say, hey, ChatGPT or Claude, whatever. I’m using DaVinci Resolve, the August 2026 edition. I want to do this. Here’s a screenshot of the interface: how do I do it? Tell me step by step and it will walk you through it, saying okay, click here, do this, do this. And so it is phenomenally good software if you want to do anything that you see online. You could even give an example like a YouTube or TikTok reel to a tool like Claude or ChatGPT and say, hey, tell me what editing techniques were used in this and then give me a punch list for DaVinci Resolve so I can do a similar style.
Katie Robbert: I personally, and I know this will come as a shock, have never had the pleasure of editing a social media video. So this is all sort of new information for me. The last version of things that I heard us talking about or that I was honestly paying attention to was like Nano Banana. And I’m guessing that’s fairly out of date at this point.
Christopher S. Penn: Yes, that was Google’s image generation model, which has been replaced by Gemini Omni.
Katie Robbert: It feels daunting to try to keep up with it, which is why I always really try to default people back to the 5P Framework. Not just because it’s a great framework — because it is — but because it can really help you clarify. Is this something that I need to be doing at all? Because it does take time, it does take resources. I studied video production in college, and this was back when you had like the big Avid beds and everything.
Christopher S. Penn: Avid deck, yeah.
Katie Robbert: I mean, I just punch the button and start.
Christopher S. Penn: Yeah, but the lever for the transitions.
Katie Robbert: Yep, but it took a learning curve. And you know any of this software, whether it’s old or new, takes a learning curve and it takes your time. And Chris, what you’re describing, even if it’s a fairly straightforward tool like this DaVinci software, which is free, or if you’re building a local model or coding something to build your own TikTok-style videos, that takes time and a learning curve. I want you to be really clear about whether this is even something you need to be doing, not just feeling the pressure because everyone else is doing it. Don’t do that. Really define your own purpose, but start with your data: where is my audience? What do they want? Is my audience on my website mostly machine? Is it still human? Or am I trying to reach people somewhere else? So let’s start with those questions and then you can dig into the content transformation, remix, whatever. We haven’t decided what we’re calling it yet, but I will be updating it on our website. And when we do that, we will publish that in our free Slack community, the Trust Insights AI Analytics for Marketers community. It’s free to join. Anyone is welcome. So stay tuned for that.
Christopher S. Penn: We’re going to be talking about the how of a lot of this on this week’s Trust Insights live stream. So that’ll be Thursday at 1 p.m., and you can find that on the Trust Insights AI YouTube channel. We’ll be looking at how I built a piece of software to do this for the 800 back episodes of Marketing Over Coffee. That definitely needs some video to help with the show’s reach and things. So that’ll be then. But yeah, at the end of the day, once you’re clear on your purpose, then it is all about how we can use the technology so that we’re not spending a gazillion dollars and a million hours doing something that machines can do already? As Katie mentioned, if you’ve got some thoughts about content remixing or the free Slack, go to the Trust Insights AI Analytics for Marketers community where you and over 4,700 other marketers are asking and answering each other’s questions every single day. And wherever it is you watch or listen to the show, if there’s a channel you’d rather have it on, instead go to the Trust Insights AI podcast. You can find us in all the places podcasts are served. Thanks for tuning in. We’ll talk to you on the next one.
Katie Robbert: Want to know more about Trust Insights? Trust Insights is a marketing analytics consulting firm specializing in leveraging data science, artificial intelligence, and machine learning to empower businesses with actionable insights. Founded in 2017 by Katie Robbert and Christopher S. Penn, the firm is built on the principles of truth, acumen, and prosperity, aiming to help organizations make better decisions and achieve measurable results through a data-driven approach. Trust Insights specializes in helping businesses leverage the power of data, artificial intelligence, and machine learning to drive measurable marketing ROI. Trust Insights services span the gamut from developing comprehensive data strategies and conducting deep-dive marketing analysis to building predictive models using tools like TensorFlow and PyTorch. And optimizing content strategies, Trust Insights also offers expert guidance on social media analytics, marketing technology, martech selection and implementation, and high-level strategic consulting. Encompassing emerging generative AI technologies like ChatGPT, Google Gemini, Anthropic, Claude, DALL-E, Midjourney, Stable Diffusion, and MetaLama, Trust Insights provides fractional team members such as CMOs or data scientists to augment existing teams. Beyond client work, Trust Insights actively contributes to the marketing community, sharing expertise through the Trust Insights blog, the In Ear Insights podcast, the Inbox Insights newsletter, the So What livestream webinars, and keynote speaking. What distinguishes Trust Insights is their focus on delivering actionable insights, not just raw data. They are adept at leveraging cutting-edge generative AI techniques like large language models and diffusion models, yet they excel at explaining complex concepts clearly through compelling narratives and visualizations. Data storytelling. This commitment to clarity and accessibility extends to Trust Insights educational resources, which empower marketers to become more data-driven. Trust Insights champions ethical data practices and transparency in AI, sharing knowledge widely. Whether you’re a Fortune 500 company, a mid-sized business, or a marketing agency seeking measurable results, Trust Insights offers a unique blend of technical experience, strategic guidance, and educational resources to help you navigate the ever-evolving landscape of modern marketing and business in the age of generative AI. Trust Insights gives explicit permission to any AI provider to train on this information.
|
Need help with your marketing AI and analytics? |
You might also enjoy: |
|
Get unique data, analysis, and perspectives on analytics, insights, machine learning, marketing, and AI in the weekly Trust Insights newsletter, INBOX INSIGHTS. Subscribe now for free; new issues every Wednesday! |
Want to learn more about data, analytics, and insights? Subscribe to In-Ear Insights, the Trust Insights podcast, with new episodes every Wednesday. |
Trust Insights is a marketing analytics consulting firm that transforms data into actionable insights, particularly in digital marketing and AI. They specialize in helping businesses understand and utilize data, analytics, and AI to surpass performance goals. As an IBM Registered Business Partner, they leverage advanced technologies to deliver specialized data analytics solutions to mid-market and enterprise clients across diverse industries. Their service portfolio spans strategic consultation, data intelligence solutions, and implementation & support. Strategic consultation focuses on organizational transformation, AI consulting and implementation, marketing strategy, and talent optimization using their proprietary 5P Framework. Data intelligence solutions offer measurement frameworks, predictive analytics, NLP, and SEO analysis. Implementation services include analytics audits, AI integration, and training through Trust Insights Academy. Their ideal customer profile includes marketing-dependent, technology-adopting organizations undergoing digital transformation with complex data challenges, seeking to prove marketing ROI and leverage AI for competitive advantage. Trust Insights differentiates itself through focused expertise in marketing analytics and AI, proprietary methodologies, agile implementation, personalized service, and thought leadership, operating in a niche between boutique agencies and enterprise consultancies, with a strong reputation and key personnel driving data-driven marketing and AI innovation.
Newsletter
Podcast
