This week, Cassandra Gaston, head of Marketing at Broadleaf, sat down with Brad Buhl, CRO, to discuss the blog creation process (leveraging GenAI) that is in the final stages of producing a whopping 100 blogs in 2025!

Listen in to hear how Broadleaf created a 179% lift with organic traffic, as well as a 344% lift with AI referrals year-over-year using their new content strategy in this Behind the Curtain look into Broadleaf - how we do what we do.

Transcript:

Brad: Welcome to the Behind the Curtain series, taking you inside ‘How We Do What We Do at Broadleaf Commerce’. Today, I am joined by Cassandra Gaston, who runs all things marketing at Broadleaf and has been at Broadleaf for nine years now. There are lots we could discuss.

We could discuss branding, product launches, and case studies. We could talk about events, whether it be virtual or in person, but 2025 for us has been the year of content. So, Cassie, how many blogs are we closing in on for just this year? 

Cassie: A whopping 100. 

Brad: 100 blogs. Awesome. And so what, have you seen the results from that work from the close to a hundred blogs now? 

Cassie: So not only have we seen a doubling in our website traffic users, but year over year, we've also seen a 179% lift with organic traffic, as well as a 344% lift with AI referrals utilizing GEO, which, for those of you who don't know, generative engine optimization, I personally think has been a key element.

For both our organic traffic and referrals, I think that these 100 almost blogs are training AIs to not only trust us, but also to keep coming back to our website to seek out information.

Brad: That's good. So today and behind the curtain, we're gonna discuss how Cassie was able to put together a system to crank out these two quality blogs a week at Broadleaf.

And I say quality because there's so much, just I think everybody says AI slop, and that's not what we're talking about today. So it's two quality blogs a week at Broadleaf, and how you can do the same. But first, Cassie, before we get into it, I want you to give us the three main takeaways that you wanna make sure that everyone who's listening today comes away with.

Cassie: Okay. So my main takeaways that I want everyone to leave with today are gonna be one, taking inventory of your assets. Two, how to use different AIs for their appropriate tasks, because not all AIs are created equal. And three, knowing where to use AI and where to rely on yourself. 

Brad: Alright, so assets, different AIs, and yourself.

So let's start, though, and back up. Start from the beginning. In the summer of 2024, we completed a website design. As a part of that website migration, how many blog articles did we bring over to the current site that we've got now? 

Cassie: So we had about 176 blogs before we migrated to the website.

After a refinement process, we decided to migrate a little under half of them to the new website and retire the rest. I think it was 72 blogs we brought over. 

Brad: Okay, so we've already done more than that this year. 

Cassie: Yes. 

Brad: And blogging. 

Cassie: Yes. 

Brad: So, migrated about half the content from 2012 to 2024. And in the past year, we've just about doubled that amount of migrated content. What about those who didn't make the migration? 

Cassie: So the ones that didn't make the migration, we are keeping on ice in case we ever wanna do updates or we simply run out of ideas and wanna mine from our own content.

Brad: Okay. Awesome. So walk us through how those first, I think it was just under a hundred blogs were created at Broadleaf, so in the first 12-year period, what did that process look like? 

Cassie: In the past, we were definitely at the mercy of bandwidth.

First and foremost, our company is made up of 95% developers, and as you can imagine, they're all very busy, so they rarely have time to create blogs from scratch on their own. Marketing would request blogs, and we would do interviews to use as a basis of material ourselves. After long editing cycles, we would get rid of a lot of the copy, which made it very ineffective for both our developer time and our time as well.

The final product would then be run through SEMrush to do SEO checks and Grammarly for grammar before going back to the dev. That was going to be the author for final approval, which again, would take weeks because they have their own things they need to do for their sprints, and this is going to take priority once all of those things are done.

Once the copy was approved, we then had to get the right photo that we could use, and we had to find photos that were perfect, that didn't have any copyright issues. And then after that, the publishing and promoting. 

Brad: Okay. So I remember at one point we would have different all-hands meetings, and then afterwards it's okay, we're giving you 30 minutes, write down the last project that you did that was cool.

And then we would take all that from everybody that did anything in that 30-minute block, and then we'd have to sift through, and then that refinement process, even when it was like, can we try to copyright this, help these bullets become an article that was, I don't think that ever was very useful. No. For us, trying to like, write on behalf of Debs, who were like, no, you're not technical. 

Cassie: Yeah. I remember there was one developer I worked with in an interview where we had a very specific topic in mind, and we were recording the whole thing. I was like, Can you break down what this part does? He spent about 30 minutes of that hour starting the example, realizing it was too technical, and then trying to warp his own example and retell it to me again and again, it made sense that I could write about it.

And by the time that we got to the end of the hour, I understood one part of the process I needed to write about. 

Brad: Do you remember in that case, did that blog ever get published? 

Cassie: No, it did not. 

Brad: Okay. So a lot of wheels turned, right? Yes. But even in that old process, I remember we tested Gen AI. There were a couple of different tools, even an earlier version of Google's Gemini, that just put computer-sounding text. It was like, this is just making stuff up.

Even when I would put in a very detailed prompt, and I know we tested this quite a bit, and I would even point to my previous blogs, but here's what I'm thinking and what I'd like to draw conclusions from. Can you help me with some copy and make it look like what I've written in the past?

Cassie: Even now, creating copy that is both accurate and doesn't sound like AI slop is a bit of a process. It takes a lot of trial and error to get down a process for creating that content as well as refining it. Which is why I said earlier, not all AI is created equal. Because in my experience, one AI is great for finding accurate data for you to write about is not going to be great with the refinement process, tone edits, and things like that. 

Brad: What has changed to make this year so different for us? 

Cassie: We can look around and see what tools are good at what. It took some time after the ChatGPT took off at the beginning of 2023, but since then, multiple tools have hit the market with their own strengths and weaknesses. In my experience, I find that using one AI for fact-checking and accuracy to build the outline of a text is great and then using a second AI for the refinement and tone of the blog has worked best for me. 

Brad: That's good stuff. The landscape is changing quickly. We'll try different tools for a season. We'll see what works. We'll go on a short-term basis with a tool, which we're doing with a couple right now, and then we'll iterate.

So how do you determine what you want to test and when you want to test it? 

Cassie: So, for me, I want a tool that's gonna make me, as a marketer, the most competitive at the best price point. For example, we used a meeting recorder called Otter AI. That was a great tool for a season. And provided some great insights from recorded meetings and interviews.

But over time, Google Gemini got good at providing the same content and context, so we didn't need to pay for that additional tool anymore. With Google Gemini, they came out with the ability to create gems with a specific purpose and a knowledge base that you create yourself. I have made many gems using our internal and external developer resources, and that has pretty much streamlined not only my ability to create technical blog material, but also reduced errors and shortened the time spent by Broadleaf developers who review these blogs before they're published.

I also learned from my husband about an AI that was created as a copy creation tool called Claude. And it has become my Gen AI choice for refining the tone and the structure of the blog post. It's really good at catching formulaic language, repetitive speech patterns, and the things that I pick up in Grammarly are going to be a lot lower than, say, ChatGPT, or if I just took the gem first draft of the outline.

I'm constantly on the lookout for new tools, so there isn't a when for me so much as it is a what?

Brad: Okay. Perfect. So I think that fits your second takeaway for today of how do you utilize different AIs for just appropriate tasks. What about the first takeaway you said, taking inventory of your assets? What did you mean by that? What is that takeaway that you're looking for people to take? 

Cassie: So when it comes to content creation, it's really important that you look through what you already have for inspiration. I took inventory of what I have as both internal and external resources that are interesting to our target audience.

We have developer documentation, and we have a ton of different data sheets. We have case studies, and we have blogs that have been put on ice. We have blogs written years ago that still get great traffic. Once I put all those things together, it's a matter of prioritizing. What do I think is going to be the most interesting thing to put out next? Are there any new ideas that we can make of that?

Brad: That's good. So that reminds me, we've got all these assets. You're taking a look at how we can reuse these, right? Or remarket these. How do we capitalize on trends that are going on? How do you look for that?

Especially, we're in commerce. Agentic commerce is obviously hot right now. 

But what are other things that you're looking for that might be more like long tail? How are you determining what the longer, more in-depth kind of question is going to be around different topics?

Where does that come from? 

Cassie: So for me, it's a couple of things. I look at our Google search console to see what the long-form questions that are being asked and led to us are, and then I reverse engineer it from there. When I look at other blogs that competitors are putting out, are there any trends there?

Are there things that are being consistently brought up over and over again?

So if I see a topic that's being talked about multiple times, then I'm like, what questions are they basically asking here? And then I look within our resources and ask, okay, so that's the question. What is our answer to that question?

Brad: Now I wanna talk about how you manage me, 'cause that's fun. When I write blogs, I'll do my first draft. I've got a very specific style and I want it to be that style. That's how I write.

Cassie: That's your signature.

Brad: Yeah. Can you just talk listeners through what it's like to edit? A Brad Buhl blog.

Cassie: So, a Brad Buhl blog, there's gonna be a lot of alliteration, there's gonna be a lot of fun examples. The language ironically enough, when I run your blogs against my AI detection gems, and I use multiple AIs just to see, like Grammarly, for instance, if you have Grammarly Premium, they tell you how much of this is plagiarized, how much of this sounds AI, and then I'll use Claude and our gems. I just average those scores. And across all of those, your blogs that are 100% human-written will score extremely high for AI language.

And the way that I manage that is I trim the fat where I'm like, okay, we have five really fun moments of alliteration. We're gonna get that down to two. We've got this many bulleted lists. We're gonna cut that in half. I first ran it through the AI. How much formulaic speech are you finding? How many repetitive speech patterns are you finding? Point them out to me and then clean 'em up. Don't change anything else.

Just wherever there's formulaic language, cut it and then give me that.

I get rid of whatever I think isn't necessary and then feed it back through. How does this sound? I tell it to score it and give it a rating for AI detection and target audience value. I go back and forth in that process until I'm happy with it, and then I send it to you.

Brad: I think it speaks to the whole like, okay, the AI models are gonna change. We've seen them change. For discoverability, just as long as it's human, that's what we ultimately resort to, right?

We're constantly looking at here's what. This tool says on SEO versus GEO, and how good is its scoring? How differentiated is it, and we care a lot about differentiation. We don't wanna sound like other people.

Cassie: Yes. 

Brad: Ultimately, I think we just go back to, is it human? 

Cassie: Yes. 

Brad: That's the thing I wanted to get into next, because there was the whole like. So I like writing, right And so I can write first and then have AI refined, like you said. 

Cassie: Yeah. 

Brad: I think when you're working with most of our developers, though, who like to write code, they don't like to write words, right? Yes. What's your process like with them? I know you'll do interviews with people. And then try to turn that into a post and still make it like their voice. With Gen AI, it does not make it a computer.

Cassie: So for me, when I write blog posts that are digital merchandising or things that are in my voice, that's super easy to do.

I'm talking about things within my wheelhouse, and I can just use AI as. A refinement point, an editing tool, or if I get halfway through it and I have writer's block, with developer content and super technical blogs. That's not my wheelhouse, and there are not enough interview hours in the world to make it my wheelhouse.

So, for me, I use the tool to bridge the gap. This is where Gemini gems specifically have come into effect. I have downloaded and compiled all of our Broadleaf developer documentation and made it into super PDFs that I then turned into gems. I will ask a gem to write about the specific thing within that subset.

Basically, what I do is go through my Google Analytics and look at what pages in our developer documentation are seeing a lot of traffic at any given time, and then I go to that page, and I read that page, and I'm like, okay, what things on this page stand out to me?

Then I go to the gem that is responsible for the developer documentation, and I ask them. What are the top three compelling elements of this developer documentation page? Because what I find interesting about it and what is actually interesting about it might not be the same thing, because I'm not a developer.

Once I get verification of what is interesting about that, I then say, Okay, I would like to draft a blog for this specific target audience. I want it to be this many words. Give me your first draft. It gives me a draft. I then take that very technically accurate, but incredibly AI-sounding blog.

And I run it over to Claude, and I'm like, Claude, let's make it sound more human. We don't want this to change the accuracy, but we wanna change the tone. This is the tone we're going for. Here is an example of a blog written by that person. How can we change it?

What's missing? And then we go through one or two refinements of that blog before I contact the person who is responsible for that specific feature. I ask them to review it and edit it once. Is it accurate? Two, is it interesting? Three. Is it worth publishing? Because I've gone through all of those steps already, verifying the accuracy against our documentation, doing the refinement to make sure that it reads well, and that it doesn't sound like AI slop, once it gets to the developer that's going to do the final pass over it, they really don't have to spend more than 30 minutes to an hour on it, making their edits, and sending it back to me for publishing.

And that frees up a lot of time for them while also guaranteeing that we're putting out content that is technically accurate and interesting for people coming to our website. 

Brad: That's great. And now I know when you started the process, though, multiple developers who'd be like, That's not even me. Delete. 

Cassie: Yes. 

Brad: What do you do to get past that? 

Cassie: So there are two examples that come to mind, and they actually helped me refine my process because at first I was just using the Gemini Gems because I was like, Oh, I can make my own knowledge base, and I spent a lot of time making that knowledge base.

It gives you the ability to give it a specific prompt, such as this is your job as a gem. Use your knowledge base to do this. I would use that gym to write the blogs, and then I would send them to the developers. And a developer was like, No offense, but this reads like AI slop, I do not want my name on this. After some pouting on Slack, I was like, fair enough. Okay, let's refine the process. I need to figure out how to make these blogs stay this accurate, but not ever get this response from another developer again. Not only have I wasted my time, but I've wasted their time; People are busy.

I was talking to my husband about this problem, and he was like, Why don't you try using an AI specifically made for copy creation? I've heard about this AI called Claude, and I was like, say less. I'm gonna check it out. Whenever the next blog that I sent out to a developer using that new part of my process, I noticed they had fewer edits. And the edits that they did have were the same edits that I had been getting across the board from multiple developers, which was, oh yeah, this sounds great, It's accurate, a little repetitive. I'm noticing a lot of the same things. I'm noticing too many bullet lists, and I'm like, these are things that I'm always going to ask the AI when I'm running it through, look out for these specific things, and so it pretty much just became a collaborative process of refining the process and each iteration noticing the same feedback that I get across the board. Adding new tools and taking constructive criticism well has gotten me to a point where now I'm sending people blogs, and they are really editing very minor things. 

Brad: Let me go through the process, then. The human AI process. It starts with an idea. And that's a human process of looking at all of the things. 

Cassie: Yes. 

Brad: It's looking at what our SEO is looking at, what's in the market. 

Cassie: Yes. 

Brad: So that human decides this is really interesting. And then you could see what our assets are that we've got today? To your first point, hey, leverage your assets. Has somebody already written a technical document on this thing? 

If so, then let's leverage that and anything else that we know about the features or benefits of this thing. So that's when you bring in the initial. Ai. It sounds like you might use one AI tool to do a bit more research. 

Cassie: Yes. 

Brad: Take it with all the data we have to create something. 

Cassie: Yes. 

Brad: Then that creation goes to a human.

Cassie: Yes.

Brad: To say, Hey, what do you think? Try to humanize and contextualize to save as much time as possible. 

Cassie: Exactly. 

Brad: And then that human puts their own spin on it, right? Hey, can you refine this? Or they're refining it themselves.  And then you go through that loop however many times, doing one final check to make sure, with all the edits. This doesn't sound too fabricated, right?

Cassie: Yeah. 

Brad: And then the last thing is the image, right? And so before we talked about that, hey, we used to, almost always use, I think it was like Flickr and Shutterstock and stuff.

What does that look like now?

Cassie: So that's also a mixed bag because sometimes with things like Adobe Firefly or Google Gemini, you can create an image that is great for the blog, put in a prompt or two, and you get what you need, and you go about your day. Other times, you are still gonna have to lean on Shutterstock or a royalty-free photo platform because when it comes to creating images with AI, it's come a long way, but it also still has a long way to go.

So it does feel like a bit of a roulette situation where sometimes you just put in a prompt and it's awesome. It gives you four images. The third one looks great. That's the one we're using. And then other times it'll be like, what about the prompt that I put in made it give me this spaghetti-looking picture?

So it's very dependent on the blog topic. The other day, we did a blog called Real Microservice, or 'True Microservices versus Fake Microservices. How to Tell the Difference.’ I was able to put that prompt in Google Gemini, and it gave me a cool image out of the four, but the other three were just monstrous. But I got lucky with one, and that's all I needed. But then there are other blogs, like a blog about scalability, for instance. And no matter how many prompts I put in, it's not gonna work out.

And if I go to Shutterstock, unless I wanna see a finger on a hologram. 

Brad: Oh, like minority report.

Cassie: Yes. If I want a minority report image, Shutterstock is the place to go. So then, in that instance, what did I do? I utilized my inventory. We have a scalability report that we put out a couple of years ago that had a lot of great images that humans created, and I was like, one of these images would be perfect for this blog. And so I pulled from assets that we already had and plugged it right in. 

Brad: Yeah. That's good. In terms of when you use AI? When do you rely on yourself?

We walked through that process. As an editor, I would love to know what some of the keywords or phrases are that the AI checker catches or misses, and what are some of the common things that you see where you're like, Oh, computer. 

Cassie: One thing, and I love this thing, but it is a dead giveaway. Em dashes, as soon as I see more than two, I'm like, an AI has cooked here.

Brad: Any kind of dash? 

Cassie: An em dash.

Brad: Oh, okay. 

Cassie: Specifically. 

Brad: Okay. Gotcha. 

Cassie: And then I've also seen articles, 'cause I read articles that are like, how to spot AI, and 'cause I made a Gem that was an AI detector using articles about AI detection.

One of them said the Oxford comma is a giveaway, and I refuse to believe that. That's propaganda. The Oxford comma, I use and I refuse to use anything else, but the em dash is very true. If I see more than two em dashes, I'm like, okay, this is heavily AI-sourced.

And then another thing is buzzwords. As soon as I start seeing a lot of business lingo, I'm like, okay, there's an AI here because one thing I will say about it is that they want to satisfy the user, which means that they are going to put a positive spin on everything.

It'll just be, the language will be too big for what it's talking about. And so anytime that I start seeing buzzwords that are aggrandizing, what they're saying, I'm just like, okay, that's AI.

As soon as I start seeing, for example, if a blog is talking about an issue and they're like "That isn't just a margin killer, it's a business ender." I'm like, "it's not X, it's Y." I could go cross-eyed talking about it. I'm so tired of seeing it.

Brad: That's good to know. I think about different sales things as well. You never wanna be salesy. I think about even phrases like at the end of the day, 

Cassie: Yep. 

Brad: Or synergies, 

Cassie: Right?

Brad: Yes. Like just that word synergies 

Cassie: The first sentence will give it away because it'll be in the landscape of this, something is always evolving. Every single time. It will be some iteration of, in this landscape, in this industry, in this blank, this is always changing, evolving, and rapidly moving. It's very formulaic. I see some version of that sentence everywhere. 

Brad: That's good to know. So I think what we need to do as part of an infographic that we offer to everybody listening to this, along with this podcast, is an AI bingo card.

I think that's what we need to create. Alright, so let's switch then and talk about best practices today that might not be best practices tomorrow.

Cassie: Do not rely on AI to do the thinking for you because you will run out of content at some point. If you lose the ability to critically think about what is good content to put out, that part of your brain will atrophy. 

Brad: Might work for a little bit. 

Cassie: Yes, it's not gonna work. It'll work for a little bit, but it won't work forever because what will end up happening is, I've seen this, I will not name names, but I have seen on certain websites.

They're pretty much cannibalizing their own content, and they're not really changing very much. They're not ingesting their own assets and dissecting them for new ideas. They are basically swapping hats on it and republishing it with a new name. And that's not good for SEO.

It's just a bad practice. It's lazy, and it's a bad practice. 

Brad: You would think that's the next level of the Google duplicate content problem. 

Cassie: Exactly, yes. And the algorithm is getting like Google's algorithm, and how it trusts domains is getting smarter every day.

So, while I think that AI, using AI to come up with blog topics or new ideas, is great as an aid, the second that you are using it as your full-blown crutch, you've got a problem. We already see a lot of that because in the past year and a half. AI has come such a long way in what it's able to do that people are gorging themselves on it.

And that's unsustainable. So, there are practices that I honestly don't think you should do today, but you definitely shouldn't do tomorrow. That's gonna be my first one. It is a tool, not a replacement. Use your brain, keep that part going.

Another thing I would say is. Don't trust it. You again need to check everything that it's putting out because hallucinations are real.

And for anybody that's unfamiliar, an AI hallucination is basically where they just make something up. 

Brad: Regardless of your prompt, 

Cassie: Regardless of your prompt, they just make it up and roll with it. There have been times when I've made gems and the specific knowledge base is only Broadleaf documentation. It's come up with hallucinations about Broadleaf. You've actually caught them and been like, what client story is this? And I'm like, that's a good point. I thought it was this client, but those numbers don't match. Hang on now. 

Brad: And it's said with so much confidence that it's the problem.

Cassie: Exactly. And regardless of the tool or how great it's been, you need to be able to verify it, and you should verify it. Don't just take it at its word. That again, hearkens back to the point of using your brain because they put it in fine print at the bottom of all of the screens, but it says, this model can be wrong. It can make mistakes. They mean it can make mistakes, it can do hallucinations. And if you publish something that isn't true, and you get called on it, that's embarrassing. Or even worse, if somebody comes across your website, reads your blog, and is wow, I wanna buy this product because of this line, and they bring it up in a call, and they're like, I have no idea what you're talking about.

Not only do you now look dumb, but the company doesn't look too great either, and it's probably gonna result in a one-on-one, which nobody wants. 

Brad: Yeah, it's kinda like a sales cycle when you could say a million things that are right, but the minute that you drop something that is just made up, then you can't be trusted from that point forward.

So that's really good. It's not a trust but verify. It's a don't trust in the first place. 

Cassie: Yes. 

Brad: And so what I heard was, hey, along the way, the places where AI's really helpful and likely will continue to be helpful are in the initial brainstorming, you said to get unblocked, right? Yes. You might be in a creative block, so help me get unblocked there. Then, in some of the validation, editing, and validation cleanup formulas at the end of the day for SEO and GEO stuff.

Cassie: Yeah. 

Brad: That AI should help you with. It's just the tool to help you. And then, with even some things like image creation, when it's just for promotion, it's not like the main thing. I think we wouldn't say hey, go ahead and create something like that looks like the Broadleaf admin, we would put in a picture of the Broadleaf admin if we wanted that. The other thing that I leverage a lot for now is deep research. 

Cassie: Yes. 

Brad: I love putting together, I'll give it like a two-page thing. I want you to go into here and I want you to do this, and then I want you to go here and I want you to do this. And then, depending on what you find there, I want you to go in to do this.

And it pulls everything back, and it might take half an hour or something. Now I've noticed something, and I'm wondering if this is a tell when it comes back, and it's like hyperlinking sources.

Sometimes those hyperlinks are like, the entire sentence is hyperlinked. 

Cassie: Yeah. 

Brad: And then I will see that in some blogs where it's like weird hyperlinks. Like, not just two words or a phrase, but an entire sentence is hyperlinked. 

Cassie: Yes. 

Brad: Is that an AI tell? 

Cassie: Yes. 

Brad: Yes. 

Cassie: The short answer is yes.

It depends on the AI tool that you use. Some will hyperlink to smaller subsets of words, and others will hyperlink to two or three sentences, which is an eyesore. I can say that in my entire career of reading articles with hyperlinks, I have never once before AI came on the scene.

Seeing entire sentences hyperlinked. It's always going to be within that sentence. What is the gold in there? Hyperlink that. 

Brad: Last question for you. Do you think in order to be a great editor, you've also gotta be a great writer?

Cassie: Yes. If you're talking about editing, just grammar, then no. You just need to be a rule follower and know the rules. But if you're being an editor in the sense of the way that books have editors, yes, the. The purpose of a true editor is not only to make sure that the grammar is there, that the citations are correct, all of the rule-following parts of editing, but also how does this sound tonally?

Is this cohesive? Is there too much fat here? Like, how bloated is this piece? You don't wanna completely lean it out to the point that it has no soul and it's just an info dump, but also you don't wanna give it so much color or try to make it so personable that anybody reading it is rolling their eyes and stops, a quarter of the way through.

Finding that balance of being colorful enough to keep my attention, but informative enough for me to deem it valuable. That comes from knowing how to tell a good story, and I think that comes from writing. 

Brad: So if you want to level up, it's not just figure out the tools to use and use this process; it's go write yourself.

Cassie: Yes. 

Brad: And read a lot. 

Cassie: Honestly, everybody's using AI. I've already said this, but I wanna say it again because I see on LinkedIn and a couple of other places that people like to use it on the side and then talk about it like they don't use it. And I find that incredibly frustrating because, being on the internet as much as I am, everybody's using it.

It is a tool. If you use it as more than an aid, what makes you great at what you do will atrophy. If you hand your critical thinking over to gain the convenience of AI, your talent will suffer. It's being able to integrate the tool with what you're already doing to make yourself better. That to me is the secret.

Brad: Well said. Thank you for being a host here, Cassie. That was awesome information. For anybody who is interested in following the Broadleaf Commerce blog, you can find it under our resources section. It's the Thought Leaf blog for everybody out there. Again, the three things. Take stock of your assets, know what you've got that you can leverage in your efforts. Number two, make sure that you utilize the different AIs for their specific tasks. Don't just take a one-size-fits-all approach. And number three, know where to use it and where to rely on yourself. If you really wanna level up, put the AI aside, focus on reading and writing. 

Cassie: Amen. 

Brad: Thanks, Cassie. 

Cassie: Thank you.