65 | Tim Moylan on Building With Claude Cowork

Most companies know they should use AI—but don’t know where to start.

In this episode, Audrey Chia welcomes Tim Moylan, tech leader and founder of Use Case, to break down how organizations can move from curiosity to implementation. Tim shares his “ignite phase” discovery approach for identifying high-ROI AI opportunities and explains how tools like Claude Cowork and OpenClaw enable teams to build internal systems, automate content pipelines, and deploy multi-agent workflows.

The conversation also explores AI safety considerations, personalization through context-driven prompting, and why learning by building beats endless tutorials.

If you want practical strategies to integrate AI into your marketing and operations, this episode delivers actionable insights you can apply immediately.

Join my weekly Newsletter: https://lp.closewithcopy.co/welcome 

Transcript

Audrey Chia 0:00

Hello and welcome back to the AI Market Display where we cover actionable frameworks to help you leverage AI and marketing strategies in your business. I'm Audrey Chair, your host, and today I'm excited to welcome Tim Moylan, an entrepreneur and tech leader with a deep focus on building with AI and video tech. Now, over the years, he has helped grown startups into global businesses, and he has also worked alongside big enterprise companies like Apple and contest. Tim is also the creator of the Beyond the Past AI newsletter, where he shares practical tips, tutorials, and insights to help professionals that you and I cut through the noise and actually learn how to work with ai. Tim, welcome to the show. Excited to have you.

Tim Moylan 0:45

Thanks, Audrey. It's great to be here. Nice intro. Thanks for the, uh, thanks for the comments.

Audrey Chia 0:50

Awesome. So Tim, tell us a bit more about yourself. When you first started playing with ai, what was that experience like and how far back did that go?

Tim Moylan 1:00

Geez, yeah. Well, AI has come so far in the last few years. I think my journey has just exponentially grown over the last few years. I, I mean, I got on board in the, in the early days of, of, I guess when Chat WT came out and, and as a technical leader and co-founder of, of, of multiple businesses. The idea of using AI to enhance your output has always been, you know, an exciting place for, for me to kind of test the boundaries of what's actually possible. and in the early days it was kind of just prompting back and forth and, creating documents was always, you know, fun and, and exciting and wow, I don't have to write everything from scratch. Uh, you know, good few years ago, but in the last, I mean, looking at the AI journey over the last few years since it was kind of released, the last three to six months has been an explosion of. Tech across the board that is now helping people not just generate, text. It's moved towards image generation and video generation and, and what I'm most excited now about building apps. and so the ability now for a co-founder or, or any entrepreneur who's looking to. Launch a business dramatically changes what you can output because there's so many tools accessible for you to use that can help you on your journey to building something, uh, without having to rely on, you know, developer over here and a content creator over there and image generation over there. So it just enables me to do so much more in my capacity to be able to grow and, and build things, test and measure very quickly, uh, and be able to scale up a business. with, with, I guess minimal effort or, or less, less people on board to actually get me, uh, to where I want to or to what I want to achieve.

Audrey Chia 2:38

Yes. I think a lot of listeners also realize the power of ai, and I like what you said about scaling businesses, right? Mm. So even for myself as a solo premier, I think the amount of work or client work that I have been doing has really multiplied because now I'm able to create my own team of mini AI assistance. You also are the founder of a company called Use Case, right? Tell us more about that and what kinds of real world use cases are there right now?

Tim Moylan 3:06

Yeah, absolutely. The use case is an exciting time for us. Uh, it's use case with a, with a K because these days you can't just get the word use case as a, as a, as a domain name, because every English word has been taken. So use Case was born to help people bring in and adopt AI in their companies. and as a, as a, as a, as an AI organization, we help people from all different pillars, from learning about how to use AI through to implementation strategies, through to project management of, of AI solutions, and then delivery, uh, out the other side in terms of helping them, uh, you know, implement AI in their organization. So Use Case was a, a really interesting place'cause a year ago when we founded the company, we realized that. A lot of people needed help in terms of understanding what use cases they could implement in their organization, and it's been an interesting journey as we've built the company now because we have. So many companies now know they just need to implement AI in some capacity, and they need help implementing it in various capacities, whether it's upskilling and training their internal staff, or looking at how they deploy projects out there into the, into the, uh, in the company, or whether it's a, a transformation project over a good number of months to deliver an uplift to their company in terms of revenue or productivity or whatever the outcome is associated with their organization. So it's been a great time to, to build, at the moment. And when we first started the company maybe a year ago, we did hire some internal resources to help us build like developers and operational people. And over the last three months, uh, we've now been able to. Really kind of push back on using external staff or internal staff to help deliver and look at focusing in on building agents and building, systems and operations around, uh, the tech that's out there now. and Claude has been an incredible resource for us and it helps us deliver so much more and build so much more so that we can, we can operate as a business far more efficiently.

Audrey Chia 5:08

Wow. I'm so curious to know, Tim. So you said that companies initially, they don't really know what use cases are there, and then they come to you guys for that support. How will you typically break down a process like that? Because most people won't know what they don't know. They, and from the folks that I had, chats with, they know AI is powerful, but they don't even know where to start. So

Tim Moylan 5:29

what

Audrey Chia 5:29

does that process look like?

Tim Moylan 5:32

Yeah, the discovery process is always an important one, uh, in, in any organization everyone has an outcome or a goal that they're looking to achieve. And so we need to go in there and, and understand. First of all, we do a, what we call an ignite phase with a company where we will go in and we will find out what problems and what solutions they're trying to, uh, achieve and run a mini proof of concept around. How do we take your problem and turn that into a solution to deliver an outcome that actually drives value for the organization? And that's generally just a, a mini week long project that we run, uh, inside a client. But the ignite phase for us helps us understand more about the customer, what they wanna achieve, uh, and then start looking at all the tools stack and everything that they might be using so that we can recommend, uh, a good solution to help that company scale. and so that, that's the quickest way for us to kind of understand what use cases are relevant in that organization. and, and helps us build that relationship on how do we take your company and deliver and an ROI out the other side.

Audrey Chia 6:34

Yeah, that makes total sense. So let's make it a bit more tangible for our listeners. Can you give us a company perhaps, uh, you know, an SMEA more traditional SME, right? Mm-hmm. That's trying to. To think of ways to implement ai, what should, what problems should they be looking at or how should they approach it?

Tim Moylan 6:54

Yeah, look, this is a great one. There's, there's, there's so many different routes. I think we can go down with any, uh, company looking to implement, uh, AI at the moment. and SMEI mean, the biggest thing that probably people are looking at at the moment is, is content generation around marketing, content generation around personal branding. That's a big one on LinkedIn the other day. Uh, these days with, with people wanting to implement, or sorry, bring on a better brand for themself as they represent their company. use cases are, I guess around, productivity inside organizations. That's a huge one. Uh,'cause every company at the moment knows they need to be more efficient in terms of people, systems and processes. And so productivity is a, is a, is a major one. But how could they get started? I guess to answer your question is, uh, is where do you get started? That is a good question. I, I'm a, I'm a big Claude ambassador and for me, Claude. Claude has really been able to accelerate my, my, my output of everything from the newsletter that I write, uh, in terms of just general content creation. the LinkedIn, the, my LinkedIn automation, uh, there's a lot of work that happens out in LinkedIn for me. in terms of the company we are building, we've got this mandate in, in use case, where we don't. Really subscribe to any tools if we can build them ourselves. And so, if you, if we're looking at, uh, we just recently built a, a PowerPoint presentation, slide generation deck that allows us to build our own custom branded presentations by just prompting them inside Claude code and delivering out a beautifully laid out presentation that has that PowerPoint style presentation format to it. And now we can enable ourselves and the team to create presentations together, branded, and built inside of our own ecosystem. So we don't necessarily need to subscribe to a PowerPoint license to be able to create presentations. and that's just the tip of the iceberg. I think we, we look at everything that we create internally, and we've now just built our own business operating system, which is a internal Claude. co, sorry, not cowork, open claw machine that handles, all of our CRM, all of our task management, all of our contact management, company management, and has AI under the hood, keeping us on track to make sure that we as a team are operating as efficiently as possible. And when you can create your own CRM internally that manages all of your business stack and all of your, your employees. That becomes a really powerful way of getting started with ai, as a startup, because you don't have to fall into the trap of going for these bigger subscriptions or, or starting with HubSpot or Salesforce or one of these, these CRMs and even the lightweight ones out there, they're not customized to your needs. So when you go and use AI to build your own solutions, it becomes far easier to tailor that application to exactly your purpose rather than try and fit into some out of the box solution.

Audrey Chia 10:06

Wow. I think really that element of personalization, right? Being able to build something that's for your company, not your company, trying to fit into something that's already built. I think that's a very important or powerful part of ai. Now, I know a lot of our listeners would be curious or interested to find out more about OpenClaw. I think OpenClaw is a term that has been going around. Many people have heard of it, but don't really know how to use it. So could you give us a, you know, the listeners a short. Overview of what it is, what is actually possible. And I think most importantly for a non-technical person, can they actually use it well enough?

Tim Moylan 10:41

Yeah, definitely. O OpenClaw I think is incredible. The technology that's coming out and, and it's come so far in just the three to four months that it's been released now. And so at our office, uh, we have two Mac Minis now running two different OpenClaw devices. But OpenClaw essentially is an AI agent on steroids because it has full access to your terminal in your machine, and the ability to create and run commands and act on your behalf. And so as an agent on, uh, this step, this separated Mac Mini, uh, that we have running, it can access chrome. It can access emails, we can build code with it. It can access anything that your MacBook or, or Windows laptop has access to so that we can implement processes that use certain functions of the laptop and have the ability of an agent take full proactive care over that particular piece of software. So our business operating system was built on OpenClaw, um. We've built, uh, a, a full, as I said, CRM that allows, us to manage our company every day. But OpenClaw manages that and it proactively looks for bugs. It proactively fixes things. It proactively heals itself. It proactively implements, features inside of, uh, the operating system. And on top of our operating system, we've also got a content generation pipeline that I'm currently building in OpenClaw. And the content generation pipeline is another great. Use case for our company because we need to drive content through our organization and there's multiple voices in our company. And, and so we've loaded up our tone of voice for each individual, person in the company. And then we've loaded our marketing plan, written by AI into, into OpenClaw. And so our mission control now keeps us accountable to drive that content through to our social media platforms. And with a calendar, a scheduler and a tick box on what's actually been done. Take a concept through to idea generation, asset generation that we're working on at the moment, and then publishing it out the other side. So OpenClaw has the ability to really become more than just one agent working in your organization. And we have now multiple agents, I think we're up to about nine agents in our open call machine having various different roles. That adopts the personality of so, uh, personas of whatever agent it needs to be. And so I've now got a team of nine agents sitting in my OpenClaw device and I can see when one agent is doing something, it's flashing, saying I'm working while another content generation agent might be sitting idle. And so I've got this team of agents all working together with one master agent that controls all of the subagents. And so OpenClaw gives you that ability to create virtual teams and have nine employees working for us, in one isolated OpenClaw environment.

Audrey Chia 13:45

Wow, that sounds extremely powerful. I'm sure. For, for someone who's listening for the first time, let's say for me, right, some questions would come to mind. how safe is it? That's number one. And number two, like how, how. How effective is the output? Right. Um, for example, when it comes to content, usually what I tend to notice is you might need that human eye to review if the quality of content is good. So given that you know the agents are just working, uh, how would you perceive the quality of their output?

Tim Moylan 14:16

Yeah, that's, and that's a very fair point because AI has the ability to spit you out anything. You'll never get a no answer from ai. It will always give you something. And so. The old term, garbage in and garbage out is, is very accurate when it comes to content generation especially. And so context management is the big, I think, focus this year to make sure that the content that you're generating is actually really good out the other side. And so I can go to any model now and say, give me a post to put on LinkedIn and it will give you a post to put on LinkedIn, but. with the amount of garbage that's coming out on LinkedIn and the term AI slop that everyone's probably heard, you need to make sure that the content that you're generating is on brand for your company, on brand, for your tone of voice. And so the context in is so critical. it's is more critical than the output, in my personal opinion, because you can always get the output on, on any model. But if you have a strong. Tone of voice, markdown file, something that's really personal about you telling the AI your goals around what you're trying to achieve, and making sure that all of that context is loaded up inside the pre generation. Then your, the output of what you get is gonna be far stronger if you have a few pages worth of documents. And I'm not just saying put in like a, you know, a single paragraph, I'm a marketer for an AI playbook. Give me a post. You need to say, I'm Audrey Chair, this is my brand, this is my LinkedIn profile. These are my company goals, these are my personal goals. These are some of the things that I've written in the past before. this is again, more as much context as you can write about me. Now take that and give me a LinkedIn post for this particular topic. And so that output will give you a far better generation than, uh, than just trying to get something, bulk output for the sake of it. And so that's where we really focus on our OpenClaw is more the pre generation side.'cause if you get that. Then you will get far better outputs in terms of quality.

Audrey Chia 16:23

Yeah, definitely. So even as a writer, if I, I tell my clients also, if you're using any AI tool, treat it like an intern, right? For an intern to be able to learn and understand and craft something. you need to give. Enough knowledge, and that is the context that I guess, Tim, you were talking about. So to my second question, right, how safe is it? Um, would a business be at risk by Yeah.

Tim Moylan 16:47

Deploying

Audrey Chia 16:48

something?

Tim Moylan 16:48

This? Yeah. Look, this, this is the, the hot topic at the moment. OpenClaw has so much capability in terms of what it can output. however, the advice at the moment that everyone is giving is running it on an isolated machine. Because you are giving an agent access to, to a machine to operate on your behalf. And there are no guardrails when it comes to open claw. So yes, security is definitely a concern, uh, at the moment. And, and that's why Apple has sold out of Mac Minis worldwide. There's a, a global shortage on, didn't know that because of open cloth. Yeah. It's, uh, it's becoming a really tough challenge to even just try and find a Mac Mini at the moment because everyone wants one to set up with, uh, with OpenClaw. I was reading a story in China where they're queuing outside of shops and all of the Apple stores have run out of, of all, Mac Minis. All resellers are, are putting their prices up because if they have any, they're in hot demand. It's. It's a really interesting device that people have chosen to run a OpenClaw instance on, but it's not, it's not the, it's not gonna head to mass market when everyone has to go and buy a Mac Mini to put on their office desktop. So we really, really need to get past that and I think OpenClaw is fantastic, but what Nvidia released just a few days ago was NemoClaw and NemoClaw takes the open claw capabilities and puts the guardrails around it. So that you can run it in an environment that is accessible to everyone and allows you to control to a degree the security aspect of not letting it go rogue on your laptop and sending emails or accessing your credit information or getting access to your passwords and containing it in that environment so that you can get it to operate under certain parameters and still achieve the same result. And so OpenClaw has definitely taken the, the world by storm in the last few months with what it can do, but it has largely been left up to people who wanna risk putting it on their own local machine or people who have separate hardware at the moment. And so. M NemoClaw and he just came out from, uh, from Nvidia, just, uh, just earlier this week. But, uh, it will see some really great traction in time because of its capabilities to put guardrails around it and install it on, on local hardware.

Audrey Chia 19:12

Wow. I'm so excited to see what happens next. Right. Especially when you make a tool a lot more accessible to so many other folks. And I think right now the early adopters are running and sprinting fast. Um, the second wave for folks, once they feel like it's safe enough, you know, they have an environment where they can test it. I think that's really when I think adoption will definitely increase. Tim, you also mentioned that you were a fan of Clot. Now why clot over the other tools?

Tim Moylan 19:39

Oh, Claude, I feel like Claude is the second love of my life at the moment. I, I spend all of my days on Claude. I burn all of my tokens every day on Claude. I love Claude because it's, it's capabilities are incredible to help me output and build things. And so I'm a, I'm a tech guy at heart. I'm a, I've always been in the tech space and to build and be able to develop applications and tools, that is where my heart lies at the moment. And Claude's Claude Code, is just incredible at what it can output in such a short amount of time. So the model for me is, uh, is exceptional on the development side. On a research side, I think it's also a very strong model as well. it's capabilities around deep analysis and deep research of topics is, is fantastic. it's halluc hallucination rate is fairly low and in the last month or two it's been very relevant, uh, that there are quite a. an organization that takes security seriously, and takes privacy seriously. and this whole saga in, in the US with the Pentagon and, and the, uh, and the, the chaos that was happening over there around using AI for, for military and, and weapon use, philanthropic has taken, I guess a, a more, a, a higher ground to make sure that they're not gonna use their tools for, for, for weaponry. and I think that's a great, uh, that's a great. Ethics, uh, model inside of their company. So I really appreciate that aspect to what I'm using Claude for as well. however, on the other, on the flip side, Claude is a little bit expensive to use as well. And so there's plenty of free models out there. There's print plenty of cheaper models, but I do certainly love the output for Claude and for me, the cost outweighs the output of what I can do every month in terms of. Content generation, competitor research, uh, asset generation, and for me especially, building tools.

Audrey Chia 21:33

Yes.

Tim Moylan 21:34

That's where I think Claude just shines.

Audrey Chia 21:36

Definitely, I think cloud has been really powerful and a lot of people, even for myself, I used to work on chat because if for from a marketer's perspective, initially chat had quite a lot of different functions, so it was a little bit ahead. Then once you start building out, you know, different, uh, knowledge bases, projects, then chat becomes the default. But right now, I think with clot there a lot more things are possible. Um, what is the main difference do you view between, uh, Claude chat and CLO cowork? Can you explain it to a beginner?

Tim Moylan 22:06

Yeah, definitely. And this is also another place that Anthropic really shines in, in what they've built with the Claude Desktop app, because Claude Chat is what everyone probably knows as chat, GBT. It's, it's a, it's a chat based interface where you can chat back and forth and get answers and results and very good contextual answers too because, the, the Claude model, I think is quite strong. Claude Chat is, is fantastic in what it can actually do, but it's, it's very much just a beginner friendly chat interface. Claude Cowork, which was released a good few months ago now, is excellent in terms of what it can do for you as an agent inside of a desktop application. And so when called. Stepping back just a little bit, Claude Code came out second. and Claude Code had the ability for developers to get in and develop applications and for me, just unleashed me to be able to do so much more. But Claude Cowork stepped on top of that and brings in the power of building applications under the hood, but doesn't necessarily need all the development technology to run it in an environment that's not friendly to a, a non-technical user. And so Claude Cowork has a lot of the power of Claude Code under the hood in a graphical user interface where you can just go and say, create me these assets, or do some competitor analysis on, on a company, and it will generate all of the documents and the presentations and the code and an interactive website. So that you can play around and click in into what you know, what output it's actually given you. So cowork I think, is a really excellent place to be able to use the power of Claude code without the developer technical setup. and, and the infrastructure needed to run Claude code. With the ability to create scheduled tasks and run, you know, run things on a, on a daily or a weekly basis, that will give you that proactive agent working on your behalf to deliver you research or competitor analysis or content generation topics or whatever you happen to, to need. You've got the power of scheduled tasks. and lastly, I think cowork also brings in the skills aspects to of what Claude Code was really great at doing. And so, Claude, as a general model is excellent, but when you have skills operating inside of cowork, you end up having a skill that is a highly tailored UI designer or a website auditor, or a ux um, researcher. And that skill can be adopted into your cowork conversation. To enable the output of a specialized skill to give you a, a very strong, context input so that your output is, is even more stronger. So yeah, coupled with skills, scheduled tasks and all of the asset generation for, um, for Claude code, it's all wrapped up into code work and a very nice, easy to use interface inside of core desktop. Yeah.

Audrey Chia 25:12

I think if you're a listener, and if you're listening to Tim, you're probably gonna be like, I need to try it right now. I'm missing out. If I'm not on clock cowork. Yeah, Tim. Right. I have two follow up questions. So for someone new, right, it seems like clock, cowork cowork is so powerful. Yet at the same time, for me, for example, I bought it from chat GPT. So my only dimension of looking at, this tool is chat. Yep. So then I'm not sure. What would be use cases that are better for cowork versus use cases that are better for chat? Or should someone just put their entire workflow to clo cowork? How would you make that distinction?

Tim Moylan 25:52

Yeah, that, that's a good question. I like to start, I start everything in Claude Cowork these days because Claude Cowork has so much more power under the hood to create different assets. And maybe I can share my screen and give you a, a quick example of a, um. Of ways that I would use Claude code to code work. and just just give you a quick example that might actually help explain the, the, the differences. That would be amazing. Sure. Lemme just share my screen here. So I'm just gonna share my entire screen. no worries. Whatever works best. Okay, cool. So Claude sits over here in, uh, in the Quad desktop app. So, uh, mine's, mine's loaded with all sorts of stuff. Yours will look a lot prettier if you signed in for the first time and not have a whole bunch of tasks and things that you're working on. Claude Chat is a, uh, is a, is similar to the chat two PT interface, where you can just start and say. If I'm gonna do well, the example I might give is doing some research actually, on your podcast, Audrey. I'm going to, do some analysis on where your podcast could be quite relevant, how you could potentially grow your podcast, and what audience you've got, what topics you've talked about over time, and, uh, and, and give a context of what the output is for both of these chats. I'm gonna start in Cloud Chat and say I want to know some more information about Audrey Chia AI Marketers podcast. give me info on all the latest 64 episodes. I will just run that just now for as a, as a quick example. So Claude chat is gonna be great. It's still gonna go out there onto the web. You can see that it's going out and doing some analysis through, apple.com, through Spotify. Uh, it's probably gonna be, uh, it's looking at one of your, buzz Sprout URLs. It'll probably go to YouTube as well. And it does a great job of using Ag agentic AI search to go out and find, probably the answer to what I'm, I've just prompted here. And so this will only take a, a good few seconds. and the output in called chat is always text-based, so you're still gonna get a, a good result here. Uh, it's now got enough information, it's building a little interactive widget for us, uh, and compiling up some information. So this is good. You might just take a minute. We'll just have to let it run.

Audrey Chia 28:29

And the super, if

Tim Moylan 28:29

you

riverside_tim_moylan _ mar 20, 2026 001_podcast 28:29

have

Audrey Chia 28:29

interesting place. I've never done this on my end before, so I don't know what the output is gonna look like. I'm curious to know what is it gonna generate.

Tim Moylan 28:38

There we go. And so it's doing some, some high level research here, which is, which is actually quite useful. it's taken a little bit longer. It's'cause you've done 64 episodes already. Uh, which is a, a great start. I can see you started the, the podcast in 2024, so well done for pushing it for nearly two years now. wow, that's great. Your average length of the podcast is, is, uh, 40 minutes, and that's, that's

Audrey Chia 29:04

true.

Tim Moylan 29:05

That's good. Validate the information. it's now going and doing some topic analysis on what your, podcast series have been about over these 64 episodes. And so once we actually start seeing some information here, this is actually probably a good comparison to what I can, uh, what I'll show you with Claude cowork.'cause it's, yeah, what we're seeing here is great. But um, I just know what we're gonna see in the next, the next one. And I'm excited to show you what's next. But we need these 64 to hurry up. Come on.

Audrey Chia 29:40

Yes. do you find that, um, if you were to run the same search query on, let's say chat or Gemini, do you find that the output varies a lot? Do you prefer like a specific response?

Tim Moylan 29:51

It does, yes. The outputs definitely vary. I, I'm. Yeah, the the answer is yes, but the answer is not that it's, there's not, I don't feel like one is better than the other because you can set up every model to, to give you certain outputs in certain ways, and it depends on what type of output you want as a person. Some people prefer the output from GPT and, and that's fine. I think Chapter GT has actually some really great output, but it just depends if you as an individual, like the output of GPT. Gemini, I also think has. Probably one of the deeper knowledge bases because Gemini is obviously owned by Google. Google also owns YouTube, so all of the context in YouTube has been preloaded into the Gemini models so that you get all the access to every, everything that's out there in YouTube as well. So you actually have a far wider reach when it comes to Gemini and, and, and a deeper knowledge of the question that you're trying to answer. and Claude, I think is a little bit in between. so the output really depends on, the person and I, I probably wouldn't recommend one over the other, but I'm just a fanboy of, of Claude. That's, that's just me.

Audrey Chia 31:05

Yeah. I can already tell. I think for me, right, it's also learning how to take the, the skill sets that've, uh, built with chat GPT, but then figuring out how to best implement it on clot to make sure it's flows seamlessly. But then. Gives me additional capabilities, I can see it has. Got it. Yeah.

Tim Moylan 31:22

Yeah, yeah, exactly. And so this is actually interesting analysis too on, on a chat side. It's given us, you know, all 64 of your episodes. It's given roughly a one paragraph breakdown of the, the date, the length, and, and what the actual, just general topics for about. And so this is, this is good. Let's give it a, uh, a good tick of approval if you wanted to go through all 64. They're all here. And that's what you get a, a bunch of text, and a and a decent output. So I just wanted to show you this in context for what I'm gonna generate now with Claude Cowork because this is useful. But let's move on to Cowork. If we jump over to Claude's Cowork. Um. Cowork operates in a very similar capacity, but what's, uh, what it has actual, what it has access to is working in a very specific folder on your laptop. So I am gonna bind this one into a, uh, an, uh, folder. I'm just gonna choose a folder on my desktop. I'll just choose,

Audrey Chia 32:25

so this is a folder that sits on your desktop, not in within cloud itself.

Tim Moylan 32:30

Correct, and they, and this is also the power of Claude Cowork as well, in that you can work locally on your machine and not send up any information through to. through to the wider models. So anyone who's working in financial documents or sensitive documents or company related documents, you can use the model to make sure that your documents stay on your machine and they don't get uploaded into the cloud. And this is also a really powerful, thing that Claude has been, has given us access to. In order to keep files on your machine, it can work on your files, but they never get sent off to the bigger world. So really great for local, machine work, but also for privacy and security as well. So prior to this, uh, this call, I actually went and just downloaded the transcripts from all of your 64, uh, episodes. I, I, again, I went to Claude and I just said, give me a transcript of all of Audrey's, uh, um, episodes on, on LinkedIn, uh, sorry, on, on YouTube. So I did do a bit of pre-work, and it took two minutes to download all your transcripts and I put them in a folder here. Uh, just for the sake of time, I won't go and do that again, but I'm gonna prompt, probably something similar. and this will take about five minutes. and in fact, I've, I've already done this actually, so I'll, I'll actually just walk through what I did because then we can just save a bit of time and see the output. And so if I jump over to this one here, uh. this output here, I did a similar prompt. I, I actually put a very, um, a much stronger prompt in here in that I wanted to analyze. Oops. Analyze all. I'm just trying to share this here. Here we go. Analyze all of your 64 podcast transcripts from Audrey's Marketing Playbook. That was essentially the same prompts that I, I put in the, in the chat window, but I wanted to do a few things extra in here. I wanted to first of all, enable a sheet with every episode breakdown. I wanted the episode title, the guest, the, the topic, anything you discussed, any key frameworks and models that were mentioned inside of your scripts. Uh, your podcasts and any tools and products that were mentioned as well. I wanted to understand the growth frameworks and so every framework or model that someone had mentioned across all of your episodes, I wanna know what framework that was. and I also wanted to, in sheet three, put in the tools and platforms that everyone has mentioned across all of your, platform, uh, uh, podcast episodes. I then wanna go and secondary create an interactive HTML dashboard. So that I can visually see all of the insights around everything that's happened with your marketing playbook, over the last 64 episodes. Um, and I wanna break it down by topics, and to give me all of the growth frameworks, the tools that were mentioned, um, and you can just use a modern style theme. And lastly, I wanted a growth strategy playbook deck that said how do you take the, uh, the top growth frameworks and apply this to your brand? And so I have loaded in a bit of context. I also went to, uh, Claude in chat and said, give me a brand context dump of law of, of Audrey. Um, and I just pasted in your LinkedIn, profile. And so I saved that in just a context file.'cause as we were talking about earlier, the context of who you are is very important for the output of these documents. So I've got that brand context file and I saved it in that folder that we just had a look at. up the top there. And I said, apply the top five frameworks to Audrey's brand, and her audience and her growth goals based on the insights from all of the episodes. And so, and, and I said, also, give me a, a progress tracker just so I can see where this project is actually up to. And then I ran that prompt. And that prompt took about, it took about 10 minutes. it went and analyzed all of your files. It created the project tracker document. It then launched a bunch of parallel agents to look at the 63 different episodes that you have. it then understood everything, uh, it said the transcript was complete. It then went and analyzed the data for any false positives to make sure that it understood what it was actually talking about. Um. It then went and and and said, now it's got better data quality to work with. And then it started building it. First started building the uh, Excel spreadsheet and then the H TM L dashboard and then the PowerPoint deck that I asked for in those three different tasks above. And then it said, great, all of these three deliverables are done. And so it gives me a quick output saying, here's my task analysis spreadsheet. My dashboard and the growth strategy playbook. And so the brilliant thing about Claude cowork is that although this is all in text, it has created in that folder on my desktop that we were just looking at all of these different assets that I asked it to. And so the first one is the spreadsheet analysis, which is, um, analysis spreadsheet here. So if I click on this, uh, asset that it's created. It shows me a full spreadsheet of every episode that you have, all 63 of them, what the title was, who the guest was, what the topic was about. Oops. The secondary topic, uh, the key frameworks that were mentioned by people inside of your podcast episodes around email marketing or prompt engineering. SEO. It gave me the tools that everyone mentioned inside of your, your, podcasts. Looks like everyone has mentioned chat, GPT. Some people will mention Jasper Stable diffusion, LinkedIn. And so it went and analyzed every tool that's ever been mentioned through your, through your podcasts. It looked at the key takeaway message from all of your episodes. takeaway message number two and number three. Wow. So there's a full spreadsheet of just data to start working with. And while that might look interesting, I think the next one is even more interesting where we go, okay, that's all great on a data level, but show me the dashboard. I actually wanna see this on a visual level and this is where it's actually really cool. And I'll, I might just open this in Chrome so that we can, can you still still see this?

Audrey Chia 38:52

Yes.

Tim Moylan 38:54

Now I have a full insights from the Marketers Playbook, the AI Marketers Playbook across all of your episodes. So similar to what chat gave us this, uh, it didn't give us this and it didn't give us any of the analysis around it, but we now have 38 frameworks, 56 tools, and of the topics, I can see that on a visual level. All of your topics are around ai. Sorry, a third of your topics are around AI strategy. Some roughly 10% in content creation, SEO B2B marketing, AI adoption. And so all of your topic distribution is set here. I can see the growth frameworks that most people have mentioned. a huge number are mentioning AI agent workflows over your episodes. personal branding is another big one that pops up in all of your episodes. customer avatars and personas. SEO is another. Pretty popular one that appears. And then going down the list, we see all the other topics that have been discussed, uh, over your episodes. And then lastly, not lastly, there's a few more to go, but I see on the framework side, AI agent workflows have appeared in 26 of your episodes. So again, this is just breaking it down into a little bit more of a detailed level. And if we go down just a little bit further, here are the tools that have actually been mentioned. So LinkedIn has been a very big conversation topic across. 61 times over your 64 episodes. Chat to PT 48 over your 64 episodes, YouTube, Gemini Chord. So I can see that all the AI tools and the, and the social platforms are a hot topic of conversation for you across all of your, your episodes and then the tools at a glance. this just gives you a numbers count of all of the tools mentioned across all of your different episodes. so quite interesting to see. Wow. Uh, and then also lastly, it just gives me, uh, this, that full spreadsheet, but in a nice, gooey way so I can go in and search for, I think I saw Charlie Hills was on one of your episodes, so there's Charlie's, uh, topic. he was talking about newsletter growth and media growth. Growth as well. That is, in my personal opinion, an excellent output for an AI analysis of your marketers playbook over the last 64 sec, uh, 64 episodes, and that took roughly five minutes to generate it.

Audrey Chia 41:16

I am blown away and speechless. My, my eyes were just like, what? Why is happening? I didn't even know that this was possible. Oh my goodness.

Tim Moylan 41:26

Yeah. This really doesn't take very long as well. And I mean, and there's also a growth strategy play deck off the back of this as well. So, uh, if you wanted to see the PowerPoint presentation that was actually created, it goes and creates a, a full presentation on how you could take your. Your playbook and puts it into a nice little presentation, and shows you the different frameworks, what agent workflows are being created, uh, your week one plan. some of the metrics look a bit broken here, but here's your personal branding, framework for launching your personal branding. So it created, a lot in terms of a a 90 day roadmap and 20 pages for a presentation on how you could grow your AI playbook series.

Audrey Chia 42:06

Wow, this is incredible. And, and I know it's not like the perfect output, but it gets you 70% of the way. Right, exactly. And it's exactly, exactly. So much of the heavy lifting,

Tim Moylan 42:17

yeah.

Audrey Chia 42:17

Researching, putting it together, making sense of the data. I think that has been a very incredible demonstration of code.

Tim Moylan 42:26

Thank you. I mean, and this is just scratching the surface as well. I mean, there's so many more things I, I could talk about, but yeah, this is just a, a quick demo of, of what's actually possible and, and I'll send you some of these assets too, so you can do your own analysis on your playbook.

Audrey Chia 42:39

That is amazing. So I'm curious to know, right, Tim, for somebody who is perhaps starting with cowork, I can see that your prompts are very specific. Um, you have multiple steps to it, so. What advice would you give someone who's trying to learn how to operate with cowork, given that it seems that it requires a different way of thinking, almost a lot more systematic.

Tim Moylan 43:00

That is, that's a good question actually. So there is plenty of prompts, templates out there on the web that you can start with. If you go to Google and search for, I want to, I, I need a prompt to, um, to help me do something. There will generally be either a YouTube video or someone that has started a prompt out there so you can find more detailed prompts that way. The way I like to do it though is goal orientated prompting, and so I will reverse engineer that and I'll say to Claude. My goal is to do analysis on Audrey Cheers podcast episodes. Can you please give me a prompt to analyze all of her episodes, find out what tools everyone's talking about, find out which frameworks people are talking about, and I want to create a an HTML dashboard about that. And so then Claude will spit out a prompt very similar to what I just showed you then. And so while that looks like a detailed prompt, the goal orientated reverse prompt. Is the best use case for getting a personalized prompt that will be relevant for whatever you're trying to do. Oh,

Audrey Chia 44:06

that is such a smart secret hack. I have so many use cases that I'm gonna test out after this. You just see me a lot. And then I'll have so many things created. But

Tim Moylan 44:18

your LinkedIn strategy will be changing to Claude Cowork now.

Audrey Chia 44:22

Yeah. No more chat GPT. But you, thanks so much for sharing that. I think, um, for me, myself, I've learned so much in this process. I think really seeing how you operate with. Cloud cowork, right. Um, helps so many of us understand that hey, this is the potential of it and we are probably just barely scratching the surface. It's such a powerful tool. So maybe to wrap us off, right, what is perhaps one piece of advice you would give business owners who are trying to integrate AI in their businesses?

Tim Moylan 44:53

Ooh, one bit of advice. My one bit of advice was start learning ai. You can do all the research and the reading and watching YouTube. but, and this might be a shameless plug, we actually run workshops on helping people start their AI journey and learn how to build applications. So find something, whether it's one of the workshops that we run or, or something in your location. Find something or a, uh, meetup that will actually help you to start on your journey of actually implementing AI and, and, and cut out the YouTube tutorials and all the learning aspects.'cause learning is excellent and very much needed, but nothing beats hands-on experience like we just went through and get into Claude and, and, and don't think I don't understand what I'm trying to achieve here. Just tell. Ai, your outcome, what do you want to do? And AI will guide you through the process to get whatever task you're looking at done.

Audrey Chia 45:50

Yes. I think even for me, being able to, um, continuously upskill myself, having that growth mindset and saying. I don't know everything, but I'm open enough to learn anything. Exactly. I think that mindset always helps. So, with that, thank you so much for joining us, Tim. We appreciated your time and insights now. Who should reach out to you and where can they find you?

Tim Moylan 46:13

Yeah, great. Thanks. Uh, it was really great to have me on board. I I appreciate your time, Audrey. Uh, I, I'm on LinkedIn. yeah, find me under Tim Moylan. and, and the company name is Use Case with a, with a K as well. Yeah, find us on LinkedIn and, and you'll find all of the, uh, the channels from there. Um, or go to our website on www dot use case, do ai.

Audrey Chia 46:34

Got it. You heard the man. And it's time for you to check out straight after this podcast. So thank you folks for tuning in and don't forget to hit the bill for more actionable AI marketing insights. We'll see you next week. Take care.