← The Dave Gerhardt Show (from Exit Five)
The Dave Gerhardt Show (from Exit Five)
How Marketers Are Creating High Quality Content with AI (That's Not Slop!)
00:59:38
Loading summary
Transcript91 lines
- [00:00]
Dave Gehard
All right, AI generated slop. I think it's the best thing to ever happen in marketing, actually, because it raises the bar, right? AI slop is going to kill deals, kill brand, and kill trust. Today, marketers like, we're also customers too, right? And so we have to actually put ourselves in the position of our customers and think about all the AI slop they're seeing. And it's on us to create things that actually matter, things that have meaning and impact, things that are educational, entertaining, funny, useful, specific and relevant. And, and that's everything that our sponsor, airops stands for. They're helping reshape how people discover and connect with brands. Because AI slop is not going to win. Airops is built for marketers who want to create content that sounds like their best subject matter expert, not another chatbot. This is content grounded in real sources, real insights and real information gain. Their content engineering platform helps you surface your highest value opportunities in AI search, then shows you how to actually take action on them. Not just see dashboards, not just get another recommendation or SEO report, but actually go out and execute. And this is the topic that everyone is being asked to get smarter about right now, AI search and SEO. If you care about this topic, then you want to go and check out Air Ops. They're built for you. It's airops.com exit5. You can learn more about Air Ops and what they're doing in the AI and SEO space. That's airops.com exit5. You're listening to the Dave Gehard Show.
- [01:23]
Adina
Exit.
- [01:32]
Dan
1, 2, 3, 4.
- [01:33]
Adina
Exit.
- [01:36]
Dave Gehard
Hey, it's me, Dave here. So real quick note on this one, dan. Dan's our CEO here at Exit 5. He hopped in and hosted this awesome webinar. Me, your favorite thought leader, favorite host, at least in my own mind. I was out. It was family vacation week. We were skiing with the kids, so Dan hopped in. He was a good VP of marketing
- [01:53]
back in the day too, if I, if I do say so myself.
- [01:55]
He, he's got the skill. He's not just the CEO, but he hosted this live session with Owen from Airops, Connor from LegalZoom, and Adina from WeFlow. Those are three people that have deep experience in AI and SEO and that's such a hot topic in our community right now. This is a really great conversation about how to create high quality content AI without putting out slop. They each walk through a bunch of their actual workflows that they're doing inside of their company. So it's very tactical and specific. If you need to get smarter and learn more about AI and SEO. This is a great one for you to tune into. Enjoyed this conversation. Dan, good job hosting this webinar.
- [02:30]
I guess I'll put in your performance review.
- [02:32]
Owen
Hey, everyone.
- [02:34]
Dan
Welcome to today's session. Sorry I'm a minute late. We're just getting ready backstage. I've got some awesome guests here today. I'm not Dave. That's the first update to give you. It's first order of housekeeping to do. I'm not Dave. I'm dan. I'm the CEO of Exit 5. We got some feedback that everyone was sick of Dave. They wanted someone new, so they chose me for some reason. Now Dave's on vacation. He's actually sick on vacation. So shout out. Dave, I hope you feel better. And I get the pleasure of hosting today's session about how to create high quality content with AI without creating AI slop. Which given how many people registered for this pretty popular topic. So that's exciting. Can you chat in? Hey, Joe, what's up? Can you chat in right now? So okay, people can hear me. I should have asked that right from the jump. You don't have to chat in. We're going to add a poll right now. The question is on a scale from 1 to 5, how much do you think AI is making B2B content more generic right now? And so there's a poll. It should appear I can't see it yet. So I don't know if it's up there yet. Let me go to polls. Yes, it's being published right now. There it is. Perfect. I just want to read the room. Yeah, okay. 10.
- [03:43]
Owen
Perfect.
- [03:44]
Dan
I want to get a feel for the room. We are here to help you today. Spend the next 60 minutes. We've got some great people on the ground on the front lines, Connor, Adina and Owen to share what they're working on, share some results, but I want to hear from you first. That'll help us make sure that we're catering this session to your needs because that's why we're all here today. It looks like we're getting a lot of fives and fours and threes. Good. It's a good mix. It's not like everybody's freaking out at a complete five level, which is good. We wanted to have a mix of everything, a real feel for how you're feeling and we want to hear from you throughout the session too. So before we add it in, I have an ad to read here. So today's session is brought to you by our friends over at airops. They are our sponsor. AI generated slop. It's the best thing to ever happen in marketing. You want to know why? Because it raises the bar. AI slop kills deals, it kills brand, it kills trust. And as a result, it's forcing marketers to create things that actually matter. That's what Air Ops is built for. It's a content engineering platform for marketers who want content that sounds like their best subject matter expert, not a chatbot ground in real resources, real insights, real information. And it helps you find and act on your highest value opportunities in AI search. Not just dashboards, but actual execution. AI search and SEO is a topic that everyone wants to get smarter on. Right now it's probably the hottest topic in exit 5. If you go into exit 5 or you read our newsletter or you come to our webinars like this and Earops helps you get there. So go check them out. Earops.com/5 Owen, for last week has been putting together a guide specifically for you of how to get started. He's going to walk you through it. I don't want to ruin it. I'll screw it up if I start talking about it. But how to get started this week on creating higher quality content with AI. He's going to give you that for free. It's actually@airops.com exit 5. You can go there, you can get the guide. Owen will tell you a little bit about it in a minute. And I'd also just say they have some amazing content on their website about AEO content creation. And so if you go to aops.com, go to resources, they do live sessions like this and they're great. So you can check those out as well. All right, let's add our guests, our guest experts, our guest speakers to stage now and we'll do some intros. All right, everybody, welcome.
- [05:57]
Owen
Hey, Don, thanks for having us.
- [05:58]
Dan
All right, Owen. Yeah, thanks for. Thanks for joining. Owen, I'm going to kick it over to you first. Let's do an intro and then let's start with. So we got some data here. Some plenty of people that are concerned about creating content that doesn't meet the bar. It's not going to help them connect to their audience. Let me ask you this question. Once you introduce yourself, why do you all feel there are so many brands struggling to connect with their audience via content right now? Great question.
- [06:20]
Owen
So formal introduction. I'm Owen. Despite how my name is spelled, I am VP of Growth at Air Ops. Like Dan said, He gave a great introduction. We're the first end to end content engineering platform. In short, we help you get found and stay found in AI search and Google. Great question. And then it's great to see the data flowing in from the poll. We do a lot of research ourselves on that topic of what our audience thinks, but I think it's like for creating quality content today, it's hard. Like it's never been easier to generate content and there's never been more tools to get context from. What's actually difficult is the prioritization of what actually matters, like what's going to move a needle for you and kind of like that age old adage of time or money or energy. I think teams now have to choose between or they're trying to make a trade off between content quality, volume and velocity. And I think oftentimes content quality is the hardest attribute to get right. It's also the attribute most teams are like skimping on and it's costing them results today and it'll increasingly cost results. And we'll see from Adina and Connor later how some of the best in the business are approaching it. And I have some data to share later as well on why content quality is actually one of the most important attributes you can go and focus on. So like the guide Dan referenced, everyone here today will leave with like kind of a playbook you can implement this week because I'm very much of the mindset like you want easy, quick wins. So I think content quality tldr one of the most important things you can focus on and it's probably the one that most marketers are leaving behind themselves right now.
- [08:01]
Dan
Awesome. Well, thanks for joining us, Owen. Obviously you have a ton of information here. You're sitting at the center of working with people like Connor and Adina who are on the front lines building out content in this new era. Conor, let's kick it over to you. Let's do a little intro and then if you could answer too, like why do you think brands are struggling so much to connect with our audience right now?
- [08:18]
Connor
Yeah, thanks. Thanks for having me as well. I'm Connor. I'm a senior SEO manager at LegalZoom. We do a lot of legal services, business formation, trademarks, etc. It's a great question. I view it as sort of an identity crisis for marketers. Right. Where historically we have for good marketers prided ourselves on providing value to the demographics we care about. And the way we've done that is by creating useful informational content. But now that so much of that demand is being met by engines, at least on the lower end of things. The question becomes, how do we as good marketers provide a new type of value that is incremental beyond what the engines are providing? And I think that a lot of our systems have not yet caught up to that reality. And so we're caught in this limbo of the things that we have been doing to connect with our customers are already being fulfilled somewhere else and in some ways better. And yet we haven't really figured out the processes needed to scale new value adds to the markets that we care about. And so the connection issue is real, but there are solutions. So hoping to talk about that a little bit more.
- [09:24]
Dan
Yeah, looking. Looking forward to digging in with you. And then we'll kick over to Romania, which is where Adina is. First of all. Adina, what time is it in Romania?
- [09:32]
Adina
It's actually 8:00pm Right, okay.
- [09:35]
Dan
End of day for sure. Is that a normal working hour for you this. This late in the day, or are you staying up late for us?
- [09:42]
Adina
Not really working hour, but I enjoy doing this type of meetings and webinars and anything, so that's totally fine.
- [09:50]
Dan
We appreciate you coming. So tell us about yourself and then, you know why, why you think brands are struggling to connect with their audience right now.
- [09:56]
Adina
Right. Thanks for having me first and basically agree with what Ion and Connor said. Jon, I'm sorry that I'm butchering your name. I can't always pronounce it right. This is. This is just being part of Romania. So I'm the head of AO at WeFlow. That's WeFlow, not Webflow. Because I know a lot of people get it wrong, including Google, which is why I'm actually here and why my role is to make sure LLMs and Google know who we are and don't confuse us with webflow. I think why brands are struggling right now is because a lot of people in SEO, myself included, we've been doing this old playbook for a lot of time. For example, two years ago, we were publishing, I was publishing 150 articles a month at User Pilot. And that was like a really high velocity that we needed to even get a chance to compete with everybody in serp. And now we're being bombarded with all of these changes in serp. Nobody knows what's working, what's not working. There's a ton of information out there what we should do. Add schema, add LLM packs, file, not at it. So it's a bit confusing. We don't know how to approach it. And a lot of people say, hey, okay, we have AI now, so let's just go and use AI to apply the same playbook that we did, scrape all the pages and everything. This is actually what I'm going to be talking about today. That's the problem. We need a mindset shift to be able to move from how do we rank in Google towards how do we teach and help. LLMs recommend us when our tool is the best choice.
- [11:34]
Connor
Awesome.
- [11:35]
Dan
And I think there's actually some specific information about what you're talking about in Owen's guide, which is again, we keep teasing Owen, but we'll talk about that in just a few minutes. Before we get into that and some research that Owen wants to share, I want to kick it over to Connor. Where are you focused right now, Connor? What are the areas you're most focused on? I actually think you have a couple slides and visuals you wanted to share as well to talk about where you're focused at. LegalZoom.
- [11:57]
Connor
So definitely a lot of focuses right now, but I think we're most interested in kind of paring down the areas in which we are focusing. So fewer pages overall, even to the extent that we're culling some long form content that's no longer serving our needs, and then focusing on really accelerating not only the quality of the remaining pages, but also smoothing out the flow from informational surfaces to transactional surfaces. So if we're getting cited and surfaced less often, you know, sessions are going down, then that means that the value of any individual session therefore goes up and we need to be nailing CRO and nailing that pathway to conversion whenever it happens. Awesome.
- [12:38]
Dan
Did you want to share a little bit? So one question actually, before Connor shares anything else, we can show you. It kind of depends on how technical you all want to get, but we can kind of show you. Like last week we met Connor, Owen and Adina and I actually went in through and they show me some of the flows they're building. I want to hear from you guys though. Again, like I said at the beginning, I want to make sure this is as helpful for you all as possible. Can you chat in, let us know if you want to see some of the flows that Connor and Adina have built and how they're actually using some of these input sources to create content. I need to hear yes and I need, I need confirmation because sometimes when we share stuff, people don't. Yeah, okay. We're getting a lot of yes, please. Okay, there we go. I was Waiting for that barrage of chats before we do it. Great. I want to go behind the scenes. I want to show this. It seems like people want it. Connor, do you mind sharing your screen, showing us some of the flows that you built?
- [13:23]
Connor
For sure. So I'll do the slides ahead and then we'll segue into the flow if that works for everyone.
- [13:28]
Dan
Perfect.
- [13:29]
Connor
Yeah.
- [13:29]
Dan
This topic deserves visuals. We don't do a lot of them in these sessions, but, like, we actually have people on the front lines building out content for LegalZoom, and we flow to pretty notable companies. And so I would love to just kind of peer behind the scenes here and see some of what they've been building and show all of you, of course.
- [13:45]
Connor
So I guess I want to start is that for better or worse, SEOs have gotten by for a very long time by creating decent content. Right. You know, you look at the SERP for a given keyword, you see what's already ranking, and then you skyscraper it in some way. You provide, you know, new subtopics or sections that weren't present, you adjust the format to surface answers more quickly and prominently, or you just polish up the writing quality itself. And that worked for a very long time as a form of kind of incremental improvements. And it specifically worked when our competitors were our actual competitors. They were other brands and businesses trying to rank in the same spaces as us and reach the same audiences. But with the advent of LLMs, more often than not, our competitors are the engines themselves. And if all we're doing is paraphrasing what previously existed, we're never going to beat the engines. Right? They set the rules, they have the tech. They're just better at it than we are. And also, why would they ever be incentivized to surface our content if it's just, in their eyes, a worse version of what they're already doing via ChatGPT or AI overviews. So the question then becomes, if that demand is being fulfilled by the engines, what do we do with our content that goes beyond what the engines can provide and hopefully what our actual competitors can provide? And how do we stop thinking about the definition of incremental as meaning small and start thinking of it as meaning new? Right? And the way we do that is by sitting down internally and having a hard conversation with each of our teams about what we have to offer that is unique to ourselves and unique to our businesses? For LegalZoom, this was, thankfully pretty simple. We've always been all about the human in the loop. Like, even as we like we just announced a partnership with Claude today. But the partnership is predicated upon human in the loop. Like we are the experts at the end of the day. So our answer was, let's start surfacing the voices of our attorneys, right? We have an in house law firm that's trademark specialty. We have access to a broad network of attorneys and other specialties. Let's elevate their voices in our content because it's not something the LLMs can do. But for other businesses, the conversation might be a little bit more complicated. Like maybe you do have subject matter experts that you can leverage either internal or external. Maybe you have proprietary data studies that you can externalize to add to the conversation. And maybe if you don't have either of those things, like for competitor comparison content, for instance, maybe instead of rehashing the third party pieces that have been done a million times, you start fresh, go in, go through the product flows yourself, and treat it as if you're doing it for the first time. Regardless of how you decide to do this and where you decide that, your incremental value kind of lies as a company. That's what we need to start thinking about as we produce content for LegalZoom. As I said, the question was pretty easy of the what, but the how was the blocker? Because creating quality at scale seems like an oxymoron, especially when you're dealing with lawyers, right? If you talked about having in person or zoom interviews with an attorney to source these quotes, you'd be lucky if you got three articles done a week. And with a portfolio of content as large as ours, it'd be 2035 before we had it rolled out across all of our content. So that's where automation comes in, right? Specifically something we called Project Penny. So penny for your thoughts. This was an attempt to automate the ops surrounding the creation of quality, rather than automating the quality creation itself, if that makes sense. It's very simple. And when we get under the hood, you're going to be much less impressed by this because it is so simple. But at the end of the day, what it is is you take the URL of an article, you run it through an LLM to identify five opportunities throughout that article where a quote from a subject matter expert would hit hardest. You then frame that question with positioning and context. You identify the most relevant expert from your knowledge base, from the subject matter experts to whom you have access, and you get an answer, you route the question to them. We're doing it through Slack, like Airops ties directly into Slack. It pings them with the message. Their tag says, hey, new quote available. Here's the quote or here's the question that we need answered. And then you re upload it right into your content. You could also do this via email or any other method of communication you want. What this allows us to do when run from a grid is essentially identify the opportunities, write the questions, find the experts, and get the answers for 1000 articles at once, rather than doing it piecemeal three a week for the next 10 years or whatever. This is, I think for us, the best way we can possibly leverage this right now. And what was very kind of nice to see when we first started doing this is how different the real world experience was from what the LLM suggested.
- [19:02]
Owen
Right.
- [19:03]
Connor
So a really good example is the first article we ever did. It had a section about how to name your business. And the LLM, when you ask it, it spits out go with a flashy name, go with a memorable name. And our business attorney actually said, no, it's the exact opposite. You want your base business name to be very vanilla and general. And then you file a DBA or a doing business as that's flashy, so that you can keep that nice vanilla name for filings and keep things straight. And then you can have the flexibility to change the flash whenever you want. So now I want to get into the flow itself, since you all were so eager and prepare to be not impressed by the.
- [19:42]
Dan
Yeah, we're getting, we're getting lots of good questions here, Connor. Please get into it. This is great. But I will, I will jump in after you show a little bit to maybe answer a couple of these questions. Sure.
- [19:51]
Connor
Like I said, the flow is not very impressive. And I've had this experience with my team as well where they think it's like the best thing since sliced bread. And then they get in, they're like, oh, I could do this. So the veneer comes off for me at least. And for context, I don't code. I built this all with the built in copilot. So it's just general kind of conversational prompting that allowed me to get to where I am today. So the flow itself, Project Penny, has two inputs. The first is just the URL of the article itself and then also a knowledge base containing the name, area of specialty, and slack ID of all of your experts. We're lucky that that knowledge base contains about 3,000 people right now for us, but it could be much more condensed for y'. All. It then goes that URL scrapes the article content, all with HTML, everything else it uses 5.1 to identify quote opportunities. These are fun. It always does. 5. It gives you the descriptor the article snippet in which it appears the required expertise type the suggested question itself and then frames it with the strategic value of why it chose this opportunity and then gives you priority levels. It then searches the knowledge base itself for the most relevant experts, ranks them for the top three and outputs directly into a dedicated Slack channel, tagging them. If you view here, this is actually an example of an output directly into Slack. So again this could be done via email or wherever else, but it just allows you to automate the ops aspect of this and reduce the need for manual communication. And then they would just comment on the Slack message saying I would like to claim this quote. Here's my answer and you take it from there.
- [21:41]
Dan
Hey Connor, can I, can I jump in with the questions? Sure, sorry, I'm going to jump into the question. I think it's the right time ask this one. So could Katina Katie and is asking about. So these are existing articles on the LegalZoom website. The whole flow here is you're asking your in house team to provide a quote. You automate it with Slack. You showed some stuff you're doing air ops actually be able to create the output for it.
- [22:03]
Connor
That's awesome.
- [22:04]
Dan
But what does it accomplish to update the article with a new quote? Like when people land on the page like the. I think you talked a little bit CRO before. Is it that you, you know, like if there's a quote or an updated quote, they're more likely to convert on that page. Like I guess Kanye was just kind of wondering like what is the value of updating these old articles?
- [22:23]
Connor
So it's twofold. It is CRO, right? Anytime that you showcase expertise, it adds trust, especially in a trust sensitive space like legal. That's always useful. But then it also fits exactly with Google's own definition of eat. Right? You are surfacing firsthand accounts from people who are very qualified to speak on the subject and then you're having all the proper schema associated with that quote. We actually do a quote module that has automated schema built into it to show Google like this is who's saying it, this is why it matters. And then the other thing is that we've seen this transition in the serps away from kind of devoid of personality generic content and transition more towards like the Reddits of the world where it's first person accounts speaking about their experiences with various topics. And as that type of content becomes more and more prominent on SERPs, including it in your content is a really good way to get exposure in the way that you really want.
- [23:25]
Dan
Awesome. So having the actual human with an actual quote is just way more valuable than it would be if you were pulling something from, you know, an article elsewhere or generically quoted.
- [23:34]
Connor
Right.
- [23:35]
Dan
So that's part of the value there.
- [23:37]
Connor
So in January we saw, I think the best example of this is we saw one of our pages get usurped by a Reddit thread and a subreddit called Working Moms about the cost of forming an LLC in Delaware or something. And that was just like really put a pin in just how important the first person POV and the lived experience is and the way that Google is choosing how to rank things.
- [24:02]
Dan
Another question from Don for you, Connor. How common do you get the same SMEs? So your legal team, the subject matter experts, identified for multiple opportunities? I guess. One question I have, just how many SMEs do you have at legalzoom of these attorneys? Do you have to ask these? And what is the frequency at which they're being asked to write quotes?
- [24:20]
Connor
For sure. So the reason you saw the same ones in that snippet is I actually fabricated knowledge base of experts for this. Didn't want to surface the names of our attorneys. But the number of times you get repetition is dependent upon the number of experts you have and the diversity of their expertise. Right. So I would suggest if you have quite a few, spend extra time on the front end really clarifying what their exact area of expertise is because that's going to let the LLM really hone in on the actual best person to answer the question.
- [24:52]
Dan
These LLMs are smart, right? As. Sorry, as Owen just said, having something unique to say gets you seen by LLM. So like having subject matter experts that have unique perspective in certain areas, like trademarking or you know, business creation, probably is part of the bipoc. This is fantastic. Hey, can everybody just like chat in? Let me know. Like this is the stuff you want to see, right? Like this is a very tactical audience is my understanding. People want to see behind the scenes, like actual flows. Just chat in, let me know that this is the kind of stuff. Yep. Kasey saying keep it up. This is great. Yes.
- [25:22]
Owen
Cool.
- [25:22]
Dan
This is great. Awesome.
- [25:23]
Dave Gehard
Hey, it's me, Dave. Our friends over at Customer I.O. are sponsors of today's episode. They're a really cool company that helps marketers turn first party data into engaging customer experiences across email SMS and push and they built their platform for marketers who actually care about the craft. Because marketing is a craft. It takes creativity, thought and taste. Right now, everyone thinks they're magically a
- [25:44]
marketer because they have access to AI
- [25:46]
and the result is kind of painful. More robotic emails, more noise, more bleh. AI isn't magic. It's not going to fix bad strategy or write great copy for you magically. But the best teams also aren't ignoring it. They treat AI as infrastructure. When it's built the right way, it actually makes marketing feel more human, not less. And that's what Customer IO is doing. Their AI handles repetitive work like setup, orchestration and tasks that should be automated so that you can focus on what actually matters. The craft of marketing, the strategy, the creativity. This is how good marketers are using AI right now. Not to replace thinking, but to support it.
- [26:21]
If this landed with you at all,
- [26:23]
this idea about the craft of marketing, I want you to go and check out Customer IO. It's Customer IO Exit 5. Go and check them out. Customer IO Exit 5.
- [26:34]
Today's episode is brought to you by Converter. They're an enterprise lead data management platform. If you're running marketing at a large B2B company, like many of you listening right now are, you know this problem well. Leads come in from LinkedIn, webinars, events, content, and the data is a mess because one form captures job title, another one doesn't. One says United States, another one says usa. By the time it hits your CRM, records are incomplete, fields don't match, and your routing is broken before the lead ever touches a sales rep. It's annoying. And now that everyone's plugging AI into their tech stack, bad data isn't just an inconvenience, it's actually a real liability because AI is going to scale. Whatever you feed it, feed it garbage, it's going to scale garbage. Converter is the layer that sits between your lead sources and your systems. With converter, every lead gets validated, enriched and standardized before it touches your CRM or marketing automation. This gives you clean data every single time. Companies like Microsoft, Amazon, Oracle and Stripe use converter today. So if you're dealing with this, you're not alone. And there's a great fix. You can check out Converter right now at converter IO exit 5. That's C O N V E R T R IO exit 5. Converter IO exit 5.
- [27:49]
Dan
All right, well, we've warmed him up for Adina. Connor did a great job showing some behind the scenes how he's building out these workflows using Slack, using EarOps, updating these old articles with quotes. We talked about how having something unique to say is important. LLMs. Dina, we want to kick it over to you. What are you doing at WeFlow? Can you share a little bit about some of the content creation you did? You were doing 150 articles a month for SEO back in the day. It sounds like that's a different strategy now. What's the new strategy?
- [28:17]
Adina
Yeah, for sure is. So back then, that was the only thing working. Like, you needed scale and you need a ton of content out there. Right now we just need better content. So it's not the volume, it's about the quality we put in there. So let me quickly share my screen
- [28:33]
Dan
here while she's sharing, anyone else in chat that's publishing was previously publishing 150 articles a month. Let's see if anyone can outdo Adina with that. With that amount of volume, that seems impressive. That seems high to me. But Adina's a pro, so I'm sure she knows what she was doing. No, no one's willing to share, I guess. I guess no one's going to beat you.
- [28:52]
Owen
You won.
- [28:53]
Adina
Okay, I won the cup of how many AI slope was put out there on the Internet. Thank you. Okay, so if there's one thing you. You get from my presentation is this line. So good content isn't AI generated. A lot of people think we need AI to generate content, which we do. We add context and everything. But what I want to change your mindset too, is that it's about AI assembled. So you will see this in all of my workflows and how I structure everything. It's about the granularity of putting tiny pieces together and having the patience to extract all of the pieces and get to the final output. So that even stands for if you want to build a full page head on. It's not going to be generate one page, is going to be generate tiny bits of pages and then put it all together. So let me show you how I'm doing this. I had a slide, really beautiful one. AI slop intelligence led. I'm not going to go through it because everybody knows this. What you need to keep in mind is that, yes, we do have that old playbook. You need to scratch that and you need to start thinking in what pieces of context, what pieces of intelligence do we have as a business? Because everybody has that being internal chat communication where people share knowledge between the team being QA reviews about your product Being interviews, being sales scripts, anything out there, that's intelligence that you can format and put into what you want to produce. So my approach is usually starting with the data. Now, thanks to Air Ops new mcp, it's actually easier to check all of the prompts and what are the gaps. So we have a ton of search console data where you can filter and see what are those long queries that people are searching for, but look mostly as AI questions, not necessarily queries. We pull that, we actually build a more comprehensive library of potential prompts that people might be asking and then we analyze the data. What we want to see is which prompts don't mention us, and when they mention us, do they mention us and talk about us the way we want them to talk? I'm referring to LLMs here. And all of these insights give us the context into which pages we want to refresh and which pages we actually miss. And LLMs are using our competitors. So it's a more targeted approach where Instead of doing 150 articles, we want to do granular 10, 20 pages and do them properly. So I'm going to show you an example of like where I started. This is a compare page where we're literally comparing ourselves to Gong. Gong is our top competitor. If you look here, you will see very surface level feature list, no dapp, no sources. Everybody's been doing this, hey, we're great, Gong's not or my competitor is really bad at what they're doing. And the reality is that LLMs won't trust this page because neither humans will do. And LLMs have one job, they are there on the user side. They want to give the best recommendation for them because they want the user to come back and keep using them, obviously. So what we did is we rebuilt pages like this from scratch using the intelligence LED process. And this is sort of like the output that we're getting to. So you're seeing this. I intentionally put small bits and pieces here because I wanted you to look at this not as the whole page, but at bits of information, sections that LLM actually see and extract and where they can actually get good information, quality information in depth comparison, non biased comparison. We actually have sections where, like where our competitor wins because let's be honest, they do win in some cases. So this is the output that we want to get to. It has a ton of LLM specific blocks, but it's also built with the human in the back of your mind. Because we're not just building for machines, we want the humans to have really in depth information. Right. Everybody hates when somebody tells you, oh, you should try that tool. You get on a trial and you realize they don't do what I need them to do. So that's a waste of time and a waste of resources for us as a brand if somebody wants and spends our time and end up not buying. So this is an overview of my whole process and instead of walking through this, I'm just going to take you through what we have in Air Ops because everybody wanted workflows and the technical side. So let me just see if this is going to change the screen for you or if I have to share this again. Okay. Seems to be working. So I'm going to start from the back pillar 5. This is the infrastructure. So everything we have in Air Ops, I'm not saying you have to use Air Ops. I just find it a really good tool to structure all my information and all my knowledge. So this has knowledge bases where we have a ton of data and then brand kit. This is something that we build ourselves. I just want you to look at the level of detail that we have in here because this is sort of what you need to avoid that slop. So the most important part in here, apart from the product lines. So this is everything about our tool is the content types. Because in here. So for the compare pages, I'm using something that has a template. I'm basically telling the LLM what I want from this page and then I'm giving it a sample which is added at the end because I can't edit it. This is probably a feature request for EON laser. Okay. Once I get this done, I have a very comprehensive competitor analysis because when you're doing competitor pages, be it blog, compare pages, you do need to understand your competitors. It's not just surface level stuff. So in here I'm actually extracting data from their sitemap from Captera in some cases where they had G2 manual review. So this is manual inputs that we have and product documentation. So all of that goes inside something that really looks through all of the data. I'm not going to go through every step in here because there's not a point, but it's extracting everything and it uses LLM to analyze and compact things. One thing that really worked for us is, yeah, analyze and summarize is doing stuff at a granular level and using a ton of which I don't have here. Of course, when you want to show something, it's not there.
- [35:42]
Dan
So when you started, obviously you had the gong versus weflow page, that's the before state. And that was. Was that one static page. And then in the new state, where you're obviously injecting this new level of intel and insight and you're building it with elements in mind. Is there multiple variations of this page or is it just one static page, though?
- [36:01]
Adina
It's still one static page. It's just more contextual and has more containers, more elements in it, like more sections.
- [36:09]
Dan
Sure. And how often is being updated? Like, is it constantly being updated?
- [36:13]
Adina
That's a really good question. That's the granularity that we can do with this type of work. Close. Because if I run competitor analysis, like every month, it's going to fetch new website data. So if my competitor updates their website to their help center, if there's new G2 reviews and everything, I'm going to get fresh data. So my end battle card here, you don't need to see this, you need to see the output. My end battle card is going to be even more comprehensive and with more accurate data right up to date. Then I take all of this and here is my grid for the pages that I want to do, create or refresh. I put in some info in here and when I build this one is actually taking that pillar approach that I was showing you first. If we go back to this, what my workflow is doing is actually pulling the LLM truth. It's fetching everything that we have from prompt data. So everything that LLMs are saying right now for specific prompts that we want to target, because we know, granted that that's where we didn't get the mention or our pages weren't used. It's also doing fresh live prompting with specific prompts to collect a different angle because we know LLMs will answer something differently every time you ask them. So we're fetching that information. This is scraping all of the competitors that we identified from the air ops data. It's building the competitor truth. So I want to understand why those pages that beat us, like, what do they have? What's the truth they are promoting out there, fetching airops data, understanding. So consolidating all of the LLM truth. Then I'm searching my knowledge base, my product, my sales transcript, I'm extracting tools, I'm looking at the competitors from there. So it's a ton of steps by step. This is what I said and assembled because I want to get the LLM truth, the competitor truth, and then my truth, and then I want to look at it and see, okay, where Are the gaps? Where are LLMs getting it wrong? What do I need to influence? Where do I need to actually put more pressure and add more context? What is wrong out there? And that gives me. When it's a refresh. I'm also analyzing existing page and I'm extracting the facts that still need to be there because I don't want to redo a page from scratch every time and have a completely different thing. So if I had some facts in there, I want to keep them, but I'm also going to validate if something hasn't changed in my product and those need to be updated. And then I take and map all my research to the content brief. This is an important section because as I said, we're taking the granular approach. So when I'm mapping, I'm actually getting a JSON, which is literally splitting all of my content into headings and calls them sections and then maps all of my intelligence. Hey, this type of intelligence goes really well in this section. So it's trying to structure everything where it fits best. And then when I'm writing it, I'm also taking the same approach and I'm writing it one step at a time because I'm not trying it to. I give it the full context. This is the page you're trying to create, but I want you to focus on this tiny bit and make it the best. And at the end, I also re edit everything because I notice that when I do that, it always becomes better. And then I have the content comparison just so I can see the differences in there.
- [39:47]
Dan
This is like 400, 500 level depth content engineering. So everybody, like, let's just be all. Let's admit that Connor and Adina have a level of depth here. Yeah, exactly. This is an insane level of like, thinking and like, and I mean that of course, in a very complimentary, insane level. Like, you all are at the very bleeding edge. And that's why Owen wanted to have both of you here to explain some of the cool stuff you're doing. All right, let's take a pause. We're showing workflows and we're showing code. Whenever we show code to marketers, by the way, everyone goes ding, ding, ding. Wait a minute. There's code I can look at someone's the. What is that movie? Is that from. I didn't know Hangover. The Hangover. That is Hangover. Okay. Yes. When he goes to the poker tables and he wins the money back.
- [40:26]
Connor
Yes.
- [40:27]
Dan
All right, let's take a timeout. Okay. Adina and Connor have clearly shown all of you that are here that they have next level thinking about content engineering, about building content in the age of AI, about how to not create a bunch of slop. Like clearly with both of these examples using real sources, real data. That's amazing. I want to answer some questions. I did one quick, easy question for the answer. What CMS are you using to publish?
- [40:49]
Adina
We're using webflow.
- [40:51]
Dan
Webflow. Weflow. Use webflow. That's an easy one to remember. No one's going to forget that. Okay, great. Let's go to the questions. I think there was a question earlier about the brand kit. So just like building the initial intel there about the brand, we just did this for instance at exit 5, we're using Claude. We built out like a style guide and we had Dave went through, wrote a bunch of different things. It kind of started there and now we're running our emails off of that project Idina. Where did you get started and what are the most important elements to include in a brand kit?
- [41:18]
Adina
So yeah, that's why I actually wanted to start with Step five, because I start with the brand kit and with the knowledge base. So I try to collect what pieces of intelligence do I have and I try to dump it everywhere, like dump it in a knowledge base and then say, okay, how can I leverage this to build all my brand foundations, to structure the data in there in a way that LLMs can read and understand? Because I know my brand kit is going to be used by LLMs and so it needs to be structured properly. I use markdown. That's a common practice. Everybody does that. LLMs are really good at understanding that. And I'm trying to be again, granular. I think I said this word 10 times in the past 10 minutes right now, but that's what I really do. Like, we don't want. This is a paragraph. This looks good on a billboard. You want that. So to get all the information, like, I'm lucky at least at WeFlow. We have a ton of sales transcripts actually in WeFlow because that's a conversation intelligence tool built for sales teams. So I can actually self select, manually select the conversations that I think are relevant and that have a ton of product context because we're also talking to a very technical audience. So I do need those technical details in there. I can't just put surface level stuff. So I need to say, hey, does WeFlow send that maps this object to Salesforce? How do you technically set it up? How do you. I don't know that thing But I can extract that information from sales calls where technical people actually explain it to other technical people. My job is to just fetch it and format it in a way that LLMs can also understand it. So that's what I do. That's my job and that's how you have to think about the brand kit. Don't just throw something in. Here's my product. So for example, in ours we have something product lines, full platform, this specific feature or that specific feature and everything is very detailed again. And also comparing to how does it. This is the pricing for that. How does security does not do key differentiators, competitors differentiator. All of this is coming from our sales calls. So nothing has been written by hand. Yes, we do validate the data because we want to make sure that it is accurate. But we are pulling from new cells, transcripts. We have a full master deck that we use in all of our cells which is constantly being updated. So the level of mistakes for us is quite small. We do give this. Hey, only extract from here. And again, the granularity processing sales calls, I'm getting, I'm formatting information, I'm processing something else, I'm formatting information, then I'm putting it all together. So it's baby steps to get the best details.
- [44:08]
Dan
All right, let's kick it over to Owen. Owen was going to share some research. So Owen obviously sits at the front of all of this, working at Air Ops, working with customers like Connor and Adena, going out and talking to people, learning he had some research about this. I think he can kind of help bring it home. We've seen two really great examples of going on a level of depth with the unique source and using humans like Honor is his team to create better content. Owen, what can you add to this conversation now? What have you been learning? What do you want to share with our audience here? Oh, I think you're muted.
- [44:40]
Owen
That'd be better. I have a ton to share. Just like overlaying what Adina and Connor went through there, I really loved how Adina phrased it as there's like the LLM truth, there's competitor truth and then there's like your truth.
- [44:52]
Connor
And.
- [44:52]
Owen
And I think irrespective of like what tool you're using, there's a lot of context which has been key to hitting on the quality aspect, which is what this conversation is about. So what they mentioned there was public information that's scrapable, sales calls, subject matter experts, performance data, like don't forget to add the quantitative in as well. As the qualitative. And then what Adina briefly touched on is like example outputs like you mentioned as well, Daniel, like what Dave did internally with your Claude co project, like give it what best in class looks like and it's not. Go to ChatGPT and like write me a blog on this. And here's like a paragraph. It's like lots of repetition and lots of context that you keep kind of compounding value over time. So that'll just be my like overlay to what the two demos went through. So what you need to know about like performance content. Because quality, yes, is important but you want quality plus performance. There's like a hundred bits of research we've done, so I'm not going to bring you through a hundred, I'm going to bring you through about three or four. But just taking one step back of how you can take action on content, in my head that's kind of four buckets. You can create content which is largely what we'll say like Connor was trying to do there. I think both of them then also touched on refreshing existing content and that's like a big place you can start if you want a good win this week. You don't need to go from 0 to 1. Add some quotes into existing content that you have and then just don't forget about off site. I'll get into some research in a second. But your presence in third party content is incredibly important. And then just like Connor had the example of like a Reddit or subreddit appearing above one of their pages. That is like a competitor you now have in search and chatgpt. So you want to just make sure that your brand is mentioned in those conversations. Here's some of the data. And again we were deep in code. This is deep in metrics. Just taking one step back. We take away from the left hand side is the bar for getting cited. The bar for appearing in AI search is much higher than actually just getting listed in your 10 blue links. Big things to take away. Structure matters. So like single H1 list sections, heading alignment, rich schema. If you're in SEO, if you're in content, none of this is revolutionary. I think just the takeaway is the bar gets higher. Like you need that really good structure which Adina had under the hood in her example. And then we spoke earlier about Velocity. Content refreshed quarterly gets three times the number of citations. And again leaning back into the quality aspect. If Conor can go back and hit up his subject matter experts once a month or once a quarter and Be like, hey, what's the latest on this? He's continually pushing ahead of the LLM and like that's where you want to, that's where you want to be to make sure that your brand keeps coming up. Don't forget about off site. So 85% of brand mentions in AI search come from third party content. That is content that you are not directly in control of. That's like very important to state. So the quality of your content will help you organically get picked up by other websites. And sure you can do like outreach, you can do backlinks. Like those tactics are incredibly important. But just be aware that you creating quality content is extremely valuable. We also need to make sure that's well distributed off of your website as well. This is the thing that we're all kind of like walking around right now, which is information gained content. This is like what you want to focus on. And if you want validation, oh my slide got skipped, I'll come back to it. But this is the big thing that Anthropic and OpenAI are communicating to people. So Connor mentioned EEAT is like Google's rules for making sure it's valuable content. OpenAI and Anthropic are literally communicating the same thing these days. In that you need to add the conversation is basically their framing of it yesterday as we said, being able to scrape what was above you who were direct competitors, repurpose their content in some way, shape or form. That was the tactic, that was the playbook we all use now. It's can you find frontier knowledge from your internal experts and put that into your content to actually get results. Oh, There we go. OpenAI they will do it based on highly targeted quality content. This is the most important thing that marketers can invest in. So there's your takeaway from the Google of our times. Frontier knowledge. Again, just putting a name on it. It's information that an LLM doesn't know or cannot yet answer or doesn't have that they can answer a user's question and think about it. We'll get into some examples later. But Ramp have tons of product level data from credit card transactions. They're now able to surface that on their pages and then that just makes everything they do far richer for humans and agents than what an LLM or their competitors are currently putting out. And this I will admit was like vibe coded as a graphic by me today. So we're using it, we're doing it live. But this is research we're conducting at the moment. And to give you Context, this is showing what attributes matter in the content you publish. Again, between Google and AI search. So in this lighter green, so the upper bar, this is kind of the standard to appear in ChatGPT. The lower bar is to appear in Google. The big things you need to take away is like more data points, more quotes, more internal links. That is going to increase your likelihood of getting into the AI search conversation. We actually did this test about 12 months ago. There was negligible difference between them. So now you can see when you have people like Connor and Adina pushing the boat out on getting their expertise into content that is making a big difference and then they're the ones pulling ahead. So the gap is definitely growing here. But it's your opportunity to kind of get on the boat and get ahead as well. I think just Decaybo just like legalzoom and we Flow, they're a good example of betting on this. And Valeria is like the content engineer over there. She started basically producing very similar workflows, but this time last year, Q1 of 2025 and you can see that their percentage of high intent leads via AI search referrals just like went off the charts. And I don't have Q4 numbers, but essentially it's like about 13, 14% as well right now. So just investment in this will bring through quality leads. Again, like we all use ChatGPT and Claude. We're doing lots of research and we probably go to two websites rather than five for our purchase decision. So one of the biggest things I guess every week as a question is like, hey, you talk about all this data. I have none of this data. Well, you're not in business as a company without like some unique edge, so be it. Like Connor's example of going to subject matter experts. You may have those as employees. They could be your software engineers, they could be other marketers and something like, like we flow carta rely on their gong calls or sales intelligence. Maybe you just have stuff that is sales enablement for the sales team. So use their battle cards and then if you're ramp kayak upwork, you have lots of product level data. You may need an engineer to help you know where it lives. But see if they can give you a CSV. See if they can give you something that's like very tangible, you can add into Claude or cloud code and start to get insights. But you definitely have internal data that you can use that we might take away. And then I am big for experimentation, like what can I do this week to get going so independent of any tools taking a big step back from Air Ops or whatever else you might have access to. And this is what I want to leave people with. And if you want a full guide of literally what I spent my full Sunday doing on this topic to validate it, you can scan that QR code. It'll bring you to the same page Dan mentioned earlier, but you can go to Google Search Console. You could go and drag your cursor over a Slack channel. You could connect a bunch of MCPs into Claude Rich Igbt. What I offered here was like pull your sales calls so go to gong again. Maybe you copy paste, maybe you can do it a level higher than that. But pull real questions, add them into your chatgpt, get it to cluster it into teams like what are people asking about? What are the terms they're using? And then from there you can do it manually or again ask it to scrape your webpage. Figure out where can you start to insert those quotes and data points onto pages that you have today. That's like one of the most high value things you can go do and is like the shortcut to validating a lot of what you saw today. So I wanted to continue the team of Stop the Slop. We actually I pulled up here. We want to see your name up in lights too. We were privileged to run a really cool brand campaign recently where Connor was one of our featured people on like Bush Shelter. So we're invested in making sure people win and like get seen in AI search. So all of these resources are free, they're non restrictive to air ops. You can bring them to any tool you want. So we just want to have people who win in this industry. So if you scan my QR code or you just email growth.com, i'll personally follow up with you. What you'll get is a full guide to what I just mentioned there on what you can do this week. People mentioned cohort in the chat. We kind of do two week live sessions where you dive deeper into recreating some of what Adina and Connor showed. Josh Grant is someone I really appreciate in the industry producing really good content. I'll give you a link to his recent substack article and then also if anyone wants a demo of air apps.
- [54:57]
Dan
Sure.
- [54:57]
Owen
Happy to walk you through or give you some more access than what's on our current free tier. But scan the QR code, email us. Happy to connect but we're here for lots of resources.
- [55:08]
Dan
Awesome. And then Connor and Adina we're going to drop your LinkedIn. I didn't ask you guys this in advance. We're going to do it anyways because you volunteered to be in the photo. We're going to drop your LinkedIn links profiles into chat right now and you can connect with Adina or Connor within the next 48 hours. If you have a very specific, logical follow up question for Conor. Dina, they can't answer questions forever, but certainly in the next 48 hours. Is that fair? Conor, Adena, they can ping you on LinkedIn and ask you a follow up question to what they showed you, what you showed today.
- [55:37]
Connor
Yeah. Give me a little grace on timing, but I'm a nerd about this stuff so I love to go down rabbit holes. Happy to help.
- [55:42]
Dan
Okay, no time limit for Connor. He's available forever.
- [55:45]
Connor
Ping him out of time.
- [55:48]
Dan
No, I just want to be respectful for your time. We really appreciate you spending it going into the details. And thank you for like showing behind the scenes like what's actually happening and everything. And Owen, thank you for sharing that guide. I read through the notion page last night. The step by step it really is literally go do this, then this and this. It's a lot of the. You know what, it's a 101 version of what Connor and Adina showed.
- [56:09]
Owen
Right.
- [56:10]
Dan
Which is creating that unique perspective where you're going to find that whether it be gong we flow other sales tools where you're recording and building out high quality content. So please check that out. That is a really useful guide. All right, we're going to poll right now. How would you rate today's session? Five being very valuable. One being, I like to say five being the best 60 minutes of your week or just like a very good 60 minutes of your week and then one being complete waste of time. Give us a rating if you could. That helps us with just with feedback and understanding. Feel free to chat in too if you. If you have any quick feedback here about today's session. We wanted to get as tactical and as specific. That's why we brought Connor and Adina in to show you the real stuff behind the scenes. I don't know if people are still filling out the poll, but thank you for submitting those. Tatiana. Sorry you missed a lot of the session. Definitely love the tactical eyes. Well and truly open. Thank you from Heather. Great information. I can't wait to share and rewatch. I think Lisa said she's sharing this with her boss right after she won the recording, so that's a good sign. Lots of thank yous here from everybody. So.
- [57:05]
Connor
All right.
- [57:06]
Dan
Thank you everyone for the time today. We won't take any more.
- [57:08]
Owen
Enjoy your three minutes back.
- [57:09]
Dan
Connor, Adena, Owen, thank you so much for spending time with us today.
- [57:13]
Owen
Thank you all. We'll see you later.
- [57:15]
Dan
All right, bye everybody.
- [57:20]
Dave Gehard
Hey, thanks for listening to this podcast. If you like this episode. Episode. You know what? I'm not even going to ask you to subscribe and leave a review because I don't really care about that. I have something better for you. So we've built the number one private community for B2B marketers at exit 5. And you can go and check that out. Instead of leaving a rating or review, go check it out right now on our website, exit5.com our mission at Exit 5 is to help you grow your career in B2B marketing. And there's no better place to do that than with us at exit 5. There's nearly 5,000 members now in our community. People are in there posting every day asking questions about things like marketing, planning ideas, inspiration, asking questions and getting feedback from your peers. Building your own network of marketers who are doing the same thing you are so you can have a peer group or maybe just venting about your boss when you need to get in there and get something off your chest. It's 100% free to join for seven days, so you can go and check it out risk free and then there's a small annual fee to pay if you want to become a member for the year. Go check it out. Learn more exit5.com and I will see you over there in the community.
- [58:30]
Today's episode is brought to you by Compound Growth Marketing. They're a full funnel demand generation agency that I've actually personally hired twice. That's right. Before I was a thought leader, I was an actual marketer, an operator, a VP of marketing myself. And CGM was one of the best agencies that I've ever hired. They help High Growth Cybersecurity, DevOps and enterprise software companies show up earlier in the buying journey where potential customers are actually forming opinions about which products to use. CGM is great because they offer the combination of AI, SEO, modern paid advertising strategy, and a dedicated go to market engineering team that you need today. So everything CGM does gets tracked, measured and improved over time. That means more pipeline for you. And this works because they were started by a former VP of marketing who gets this space. They really understand B2B. So if you're in search of a new agency that can help you hit the number this quarter and you need help with things like AI, SEO and paid media. You should definitely go and check out Compound Growth Marketing. I call them CGM Compound Growth Marketing. Go and check them out at compound growth marketing.com and tell them that Dave and Exit 5 sent you.