In this episode of In-Ear Insights, the Trust Insights podcast, Katie and Chris discuss the practical application of AI agents to automate mundane marketing tasks.
You will define what an AI agent is and discover how this technology performs complex, multi-step marketing operations. You will learn a simple process for creating knowledge blocks and structured recipes that guide your agents to perform repetitive work. You will identify which tools, like your content scheduler or website platform, are necessary for successful, end-to-end automation. You will understand crucial data privacy measures and essential guardrails to protect your sensitive company information when deploying new automated systems. Tune in now to see how you can permanently eliminate hours of boring work from your weekly schedule!
Watch the video here:
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, one of the things that people have said, me especially, is that 2026 is the year of the agent. The way I define an agent is it’s like a real estate agent or a travel agent or a tax agent. It’s something that just goes and does, then comes back to you and says, “Hey, boss, I’m done.”
Katie, you and I were talking before the show about there’s a bunch of mundane tasks, like, let’s write some evergreen social posts, let’s get some images together, let’s update a landing page. Let me ask you this: when you look at those tasks, do they feel repetitive to you?
Katie Robbert: Oh, 100%.
I’ve automated a little bit of it. And by that, what I mean is I have the background information about Trust Insights. I have the tone and brand guidelines for Trust Insights. So if I didn’t have those things, those would probably be the biggest lift.
And so all I’m doing is taking all of the known information and saying, okay, let’s create some content—social posts, landing pages—out of all of the requirements that I’ve already gathered, and I’m just reusing over and over again. So it’s completely repetitive.
I just don’t have that more automated repeatability where I can just push a button and say, “Go.” I still have to do the work of loading everything up into a single system, going through it piece by piece. What do I want? Am I looking at the newsletter? Am I looking at the live stream? Am I looking at this podcast? So there’s still a lot of manual that I know could be automated, and quite frankly, it’s not the best use of my time. But it’s got to get done.
Christopher S. Penn: And so my question to you is, what would it look like? We’ll leave the technology aside for the moment, but what would it look like to automate that? Would that be something where you would say, “Hey, I want to log into something, push a button, and have it spit out some stuff. I approve it, and then it just…”
Katie Robbert: Goes, yeah, that would be amazing. I would love to, let’s say on a Monday morning, because I’m always online early.
I would love to, when I get up and I’m going through everything in the background, have something running, and I can just say, “Hey, I want two evergreen posts per asset that I can schedule for this week.” You already have all of the information. Let’s go ahead and just draft those so I can take a look. Having that stuff ready to go would be so helpful versus me having to figure out where does.
It’s not all in one place right now. So that’s part of the manual process is getting the Trust Insights knowledge block, finding the right gem that has the Trust Insights tone, giving the background information on the newsletter and the background information on the podcast and so on so forth, making sure that data is up to date. As I was working through it this morning and drafting the post and the landing pages, the numbers of subscribers were wrong. That’s an easy fix, but it’s something that somebody has to know. And that’s the critical thinking part in order to update it appropriately. Those kinds of things, it all exists. It’s just a matter of getting into one place.
And so when I think about automation, there’s so much within our business that gets neglected because of these—I’m not going to call them barriers—it’s just bandwidth that if I had a more automated way, I feel like I would be able to do that much more.
Christopher S. Penn: So let’s think about this. There’s obviously a lot of systems, Claude Code, for example, and QWEN Code and stuff, the big heavy coding systems. But could you put all those requirements, all those basics into a folder on your desktop?
Katie Robbert: Oh, absolutely.
Christopher S. Penn: Okay. And if you had some help from a machine to say, “Hey, looks like you’re using our social media scheduling software, AgoraPulse. AgoraPulse has an API?”
Katie Robbert: Yep.
Christopher S. Penn: Would you feel comfortable saying to a machine, “AgoraPulse has an API. Here’s the URL for it. I ain’t going to read the documentation. You’re going to read the documentation and you’re going to come up with a way to talk to it.”
Would you then feel comfortable just logging into, say, Claude Cowork, which came out recently and is iterating rapidly? It is becoming Claude Code for non-technical people.
Katie Robbert: Yep.
Christopher S. Penn: And Monday morning, say, “Hey, Claude, good morning, it’s Monday. You know what to do.” Invoke the Monday morning skill. It goes and it reads all the stuff in those folders because you’ve written out a recipe, a process, and then it says, “Here’s this week’s social posts. What do you think?” And you say, “That looks good.” And by the way, all of the images and stuff are already stored in the folders so you don’t need to go and download them every single time. This is great. “I will go push those to the AgoraPulse system.” Would that be something that you would feel comfortable using that would not involve writing Python code after the first setup?
Katie Robbert: Oh, 100%. Because what I’m talking about is when we talk about evergreen content—and I’m not a social media manager, but we’re a small company and we all kind of do everything—this is content that’s not timely. It’s not to a specific. It only works for this quarter or it only works for this specific topic. Our newsletter is evergreen in the sense that we always want people subscribing to it. We always want people to go to TrustInsights.ai/Newsletter and get the newsletter every Wednesday.
The topic within the newsletter changes. But posting about the fact that it’s available for people to subscribe to is the evergreen part. The same is true of the podcast, we want people to go to TrustInsights.ai/TIpodcast, or we want people to join us on our live stream every Thursday at 1:00 PM Eastern, and they can go to TrustInsights.ai/YouTube. What changes is the topic that we go through each week, but the assets themselves are available either live or on demand at those URLs at all times. I just wanted to give that clarification in case I was dating myself and people don’t still use the term evergreen content.
Christopher S. Penn: Well, that makes total sense. I mean, those are the places that we want people to go. What I’m thinking about, and maybe this is something for a live stream at some point, is now that we have agentic frameworks for non-technical people, it might be worth trying to wire that up. If we think about it, of course, we’re going to use the 5Ps. What is the purpose? The purpose is to save you time and to have more things automated that really should be automated. And obviously, the performance measure of it is stop doing that thing. It’s 2 seconds on a Monday morning, or maybe 2 seconds on the first of the month. Because an agentic framework can crank out as much stuff as you have capacity for. If you buy the Claude Max plan, you can basically create 2 years worth of content all in one shot.
And so it becomes People, Process, Platform. So you’re the people. The process is writing down what you want the agent to do, knowing that it can code, knowing that it can find stuff in your inbox, in your folder that you put on your desktop, knowing that it can reference knowledge blocks. And you could even turn those into skills to say, “Trust Insights Brand Voice is now a skill.” You’ll just use that skill when you’re writing. And the platform is obviously a system, like Cowork. And given how fast it’s been adopted and how many people are using it, every provider is going to have a version of this in the next quarter. They’d be stupid if they didn’t.
That’s how I think you would approach this problem. But I think this is a solvable problem today, without buying anything new—because you’re already paying for it. Without creating anything new, because we’ve already got the brand voice, the style guide, the assets, the images. What would be the barrier other than free time to making this happen?
Katie Robbert: I think that’s really it. It’s the free time to not only set it up, but also to do a couple of rounds of QA—quality assurance. Because, as I’ve been using the Trust Insights Brand Voice gem this morning, I’m already looking at places where I could improve upon it, places where I could inject a little more personality into it, but that takes more time, that’s more maintenance, and that just makes my list longer. And so for me, it really is time. Are the knowledge blocks where I want them to be? Do I need to? This is my own personal process. And this is why I get inundated in the weeds: I start using these tools, I see where there could be improvements or there needs to be updates. So I stop what I’m doing and I start to walk backwards and start to update all of the other things, which just becomes this monster that builds on itself. And my to-do list has suddenly gotten exponentially larger. I do feel like, again, there’s probably ways to automate that. For example, send out a skill that says, “Hey, here’s the latest information on what Trust Insights does. Update all the places that exist.” That’s a very broad stroke, but that’s the kind of stuff that if I had more automation, more support to do that, I could get myself out of the weeds.
Because right now, to be completely honest, if I’m not doing it, that stuff’s not getting done. So nobody else is saying, our ideal customer profile should probably be updated for 2026. We all know it needs to be done, but guess who’s doing it? This guy with whatever limited time I have, I’m trying to carve out time to do that maintenance. And so it is 100% something I would feel comfortable handing off to automation with the caveat that I could still oversee it and make sure that things are coming out correctly so it doesn’t just black box itself and be like, “Okay, I did these 20 steps that you can no longer see, and it’s done.” And I’m like, “Well, where did it go wrong?” That’s the human intervention part that I want to make sure we don’t lose.
Christopher S. Penn: Exactly. The number 1 question that people need to ask for any of these agentic tools for figuring out, “Can I do this?” is really simple: Is there an API? If there is an API, a machine can talk to a machine, which means AgoraPulse, our social media scheduling software, has an API. Our WordPress website—our WordPress itself has an API. Gravity Forms, the form management system that we have, has an API, YouTube has an API, etc.
For example, in what you were just talking about, if you set up your API key in WordPress and gave it to Claude in Cowork and said, “Hey, Claude, you’re going to need to talk to my website. Here’s my API key. You write the code to talk to the website, but I want you to use your Explore agents to search the Trust Insights website for references to—I will call it dark data. Make me a list, make me a spreadsheet of all the references to dark data on a website, with column 1 being the URL and column 2 being the paragraph of text.” Then you could look at it and go, “Hey, Claude, every time we’ve said dark data prior to 2023, we meant something different. Go.” And using the WordPress API, change those posts or change those pages.
This is the—I hate this term because it’s such a tech bro term, but it actually works. That is the unlock for a web, for any system: to say, is there an API that I can literally open up a system? And then as long as you trust your knowledge blocks, as long as you trust your recipe, your process, the system can go and do that very manual work.
Katie Robbert: That would be amazing because you know a little bit more about my process. This morning, I was on those two systems. I was on our WordPress site, and I was on our YouTube channel. As I was drafting posts for our podcast, I went to our YouTube channel and took a screenshot of our playlist to get the topics that we’ve covered so that I could use those to update the knowledge block about the podcast, which I realized was outdated and still very focused on things like Google Analytics 4. It wasn’t really thinking about the topics we’ve been talking about in the past 6 to 12 months. I did that, and I also gave it the content from the landing page from our website about the podcast, realizing that was super out of date, but it gave enough information of, “And here’s all the places where the podcast lives that you can access it.” It was all valuable information, but it was in a few different places that I first had to bring together. And you’re saying there’s APIs for these things so that I don’t have to sit here with every other screenshot of Snagit crashing, pulling out my hair and going, “I just want to write some evergreen posts so that more people subscribe?”
Christopher S. Penn: That’s exactly what I’m saying.
Katie Robbert: Oh, my goodness.
Christopher S. Penn: And I would say, now that I think about this, what you’re describing, you wouldn’t even need to use the API for that.
Katie Robbert: Great.
Christopher S. Penn: Because a lot of today’s agentic tools have the ability to say, “I can just go search the web. I can go look at your YouTube channel and see what’s on it.” And it can just browse. It will literally fire up a browser. So you can say, “I want you to go browse our YouTube channel for the last 6 months. Or, here’s the link to our podcast on Libsyn. I want you to go browse the last 25 episodes. And here’s the knowledge block in my folder on my desktop. Update it based on what you browse and call it version 2 so that we don’t overwrite the original one.”
Katie Robbert: Oh, my goodness.
Christopher S. Penn: Yeah, that. So this is the thing that again, when we think about AI agents and agentic AI, this is where there’s so much value. Everyone’s focused on, “I’m going to make the biggest flashes.” No. You can do the boring crap with it and save yourself so much sanity, but you have to know where to get started. And the system today that I would recommend to people as of January 2026 is Claude Cowork. Because you already installed Claude on your desktop, you tell it which folder it can work in so it’s not randomly wandering all over your computer and say, “Do these things.” And it’s no different than building an SOP. It’s just building an SOP for the junior most person on your team.
Katie Robbert: Well, good news, that is my bailiwick: SOPs and process. And so, shocker, I tend to do things the exact same way every single time. That part of it: great, it needs a process done. It’s going to take me 2 seconds to write out exactly what I’m doing, how I want it done. That’s the part that I have nailed. The question I have for you, because I’ll bet this question is going up from a lot of people, is what kind of data privacy do we need to be thinking about? Because it sounds like we’re installing this third-party application on our work machines, on our laptops, and many of us keep sensitive information on our laptops—not in the cloud, not in Google Drive or SharePoint, wherever people have that shared information. Obviously, we’re saying you can only look at these things, but what is it? What do we need to be aware of? Is there a chance that these third-party systems could go rogue and be like, “Effort? I’m going to go look at everything. I’m going to look at your financials, I’m going to get your social. That photo that you have of your driver’s license that you have to upload every 3 months to keep your insurance? I’m going to grab that too.” What kind of things do we need to be aware of, and how do we protect ourselves?
Christopher S. Penn: It comes down to permissions. The Anthropic’s app—I should be very clear about this—Anthropic’s app is very good about respecting permissions. It will work within the folder you tell it and it will ask you if it needs to reference a different folder: “Can I look at this folder?” It does not do it on its own.
Claude Code. There is a special mode called Live Dangerously which basically says, “Claude, you can do whatever you want on my system.” It is not on by default. It cannot be turned on by default. You have to invoke it specifically. QWEN’s version is called YOLO. Cowork doesn’t even have that capability because they recognize just how stupidly dangerous that is.
If you are working on very sensitive data, obviously the recommendation there would be to use it in a different profile on your computer. If your Windows machine or your Mac can have different profiles, you might have an AI only profile that will have completely different directories. You won’t even be able to see your main user’s. And then if you’re really, really concerned about privacy, then I would not use a cloud-based provider at all. I would use a system like QWEN Code, which does not have telemetry to relay back to anybody what you’re doing other than actions you take, like you turned it on, you turned it off, etc. And you can download QWEN Code source and modify it to turn all the telemetry off if you want to, or just delete it out of the code base and then use a local model that has no connection to the Internet if you’re working on the most sensitive data.
Katie Robbert: Got it. I think that’s incredibly helpful because you and I, we’re very aware of data privacy and what sensitive data and protected data entails. But when I think about the average marketer—and it’s not to say that they don’t care, they do care—but it’s not top of mind because they’re just underwater trying to find any life raft to get out of the weeds and be like, “Okay, great, this is a great solution, I’m going to go ahead and stand it up.”
And data privacy tends to be an afterthought after these systems have already accessed all of your stuff. Again, it’s not that people using them don’t care, it’s just not something that they’re thinking about because we make big assumptions that these tech companies are building things to only do what they’re saying they do. And we’ve been around long enough to know that they’re trying to get all.
Christopher S. Penn: Our data exactly. The where the biggest leak for the casual user is going to be is in the web search capabilities. Because we’ve done demos on our live streams and things in the past of watching the tools do web search. If you do not provide it a secure form of web search, it will just use regular web search, and then all that stuff can be tracked back to your IP, etc. So there are ways to protect against that, and that’s a topic for another time.
Katie Robbert: All right, go ahead.
Christopher S. Penn: I think the next steps we should be doing is let’s get Claude Cowork set up maybe on a live stream and get the knowledge blocks without them being updated and say, “Let’s do this as a first test. Let’s try to update these knowledge blocks using web search tools and see what Claude Cowork can do for you.”
Katie Robbert: I was going to suggest the exact same thing because if you’re not aware, every week, every Thursday at 1:00 PM Eastern, we have our live stream, which you can catch at TrustInsights.ai/YouTube. And we walk through these very practical things, very much a how-to. And so I love the idea of using our live stream to set up Claude Cowork. Is that what it’s called?
Christopher S. Penn: That’s what it’s called, yes.
Katie Robbert: Because I feel like it’s easy for you and I to talk about theoretically, “Here’s all the stuff you should do,” but people are craving the, “Can you just show me?” And that’s what we can do on the live stream, which is what I was trying to write for social posts, full circle. “Here’s the podcast, it introduces the idea. Here’s the live stream, it’s the how-to. Here’s the newsletter. It’s the big overarching theme.” I was trying to write social posts to do all of those things, and my gosh, if I just had an agent to do it for me, I could have done other things this morning because I’ve been working on that for about 2 hours.
Christopher S. Penn: Yep. So the good news is once we do this, and once you start using this, you never do that again. That’s always the goal of automation. You solve the problem algorithmically and then you never solve it again. So that’ll be this week’s live stream.
Katie Robbert: Yes.
Christopher S. Penn: If you’ve got some thoughts about how you’re using AI agents to take care of mundane tasks, pop on by our free Slack. Go to TrustInsights.ai/analyticsformarketers, where you and over 4,500 other marketers are asking and answering each other’s questions every single week. And wherever it is that you watch or listen to the show, if there’s a channel you’d rather have it on, go to TrustInsights.ai/TIpodcast. You can find us at all the places where podcasts are served. Thanks for tuning in and we’ll talk to you on the next one.
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 and MarTech selection and implementation, and high-level strategic consulting. This encompasses emerging generative AI technologies like ChatGPT, Google Gemini, Anthropic Claude, DALL-E, Midjourney, Stable Diffusion, and Meta Llama. Trust Insights provides fractional team members such as CMO 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?* live stream, webinars, and keynote speaking.
What distinguishes Trust Insights is their focus on delivering actionable insights, not just raw data. Trust Insights 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
