"Earning the Right to Automate" | Eoin Clancy

Welcome back to the GTM Engineers School podcast, Season 2: From Hype to Operating System. Because now that we know that GTM engineering is here to stay, we now want to know where it's heading. And today we're getting into one of the biggest shifts in GTM right now, for sure, the move from SEO and content to AO and AI search and content engineering. So what happens when you stop publishing basically and actually start building systems of content at scale that compounds? And we have Owen Clancy today, the VP of Growth at AirOps, which is the only end-to-end AI search platform that a lot of you already use inside your workflows. And here's what's, what makes Owen's story super interesting, I think, for our audience, that he started himself in the trenches as a builder, as a growth engineer. He spent over 6 years at Telnyx climbing from growth engineer to head of growth. And that's where he got his first taste of actually building with AI. Before it was cool, if you'd like. And in late 2023, he also founded Build AI First, going all in on AI-driven growth strategies and community building, which is something that he carried with himself now standing up the super successful Airobs Content Engineering cohort, which we super recommend that you guys look into. And then that's where he joined Airobs and helped scale the company from $2 million to $15 million in ARR in just 11 months. So again, Owen's not just running a team. He came up through the same path that a lot of you are going on right now. Growth engineer turned growth leader with technical chops and building hands-on in the trenches, the system that his team now manage. And so today we'll go, we'll get into how content engineering is different from simple content marketing and SEO. What it takes to actually win in AI search and AEO. Spoiler, that might relate to context and what he's learned about building a proprietary system that competitors can't actually replicate because again, that's also part of what makes the rush for like how you can win in this rush of AI content. So Oren, we're pumped to have you on the show.

Thanks for having me. I'm excited to get into it.

Alrighty, then. I mean, like, let's start with some basic sensemaking, right? You know, like, we know that GTM engineering, content engineering is not new anymore. You know, like, God knows, like, we also like added, we've just added probably like context engineering and vibe coding engineer engineering this year. So, you know, there's a lot of blending between subject domain expertise and subject matter expertise as with, you know, AI enablement and workflows. But from your point of view, running the team at Airops, how have you seen, you know, like this movement change in the last year?

Yeah, really great question, Matteo. I think people are still at different stages of their own journeys or like the adoption curve. So people are still just getting into like Zapier and like those are kind of the laggards and you have people who are like blazing ahead on Cloud Code, vibe coding for days. So I think like the, the spectrum is potentially wider than ever and you're like top marketer, your top growth person, your top GTM engineer, they're actually moving further and further away from like the median. So people who are spending like days and days in Claude Code are just like developing at like a much faster rate than your average marketer. So I think there's like this gap that is emerging. I think like some of the trends I've particularly noticed with my own like usage, and I think a lot of our users and especially my team as well, um, are coming across, is that we're moving from kind of very dedicated specific tools and like out-of-the-box functionality more towards, um, like custom custom in-house built solutions. Like, everyone's go-to-market motion is going to be very different. Like, I've gone on other podcasts before just talking about our webinars or just talking about like our paid strategy. So like, we actually don't have anything really all that custom in-house today except the, the time and effort we spent on like AROPS workflows. So I think we've definitely seen this shift across, as you mentioned, Vibe Code engineering to GTM engineering to content engineering, where users are like just in a dedicated tool to now they need to like centralize centralize all of their context into either like that tool or something that they build in-house. I just think it's gone from being super like siloed to now you need to be able to pull from multiple different places and centralize it so that your process and workflow becomes a lot more streamlined.

Great. And I think like this concept of polarization of, you know, skills and/or quality, you know, like The good are gonna become, have a chance at become really, really good, very fast. If you have, if you behave like a prosumer, if you have agency, if you have clarity and stuff like that. And the mediocre are gonna become, are gonna be replaced, they're like even faster, right? And I'd love to, before we jump into the weeds of the episode, I'd love to hear from you. What are, in your opinion, the implications of these up-and-coming polarizations? Because, you know, we know that the market is moving at breakneck speed. We know that this shift is only going to accelerate, you know, like, and with this polarization, you know, like, which obviously, you know, comes with all the urgency about, you know, like the job market and, you know, like, oh, AI is coming for our jobs or You know, like, or everyone on LinkedIn seems to be doing everything at once and super well.

Yeah, there's a lot in that. Like, 2 or 3 things I'll just like pick out are, one, I don't know if people listen to like the All In podcast. It's one of my favorite things. Saturdays, spend 90 minutes going through it. But I subscribe to David Friedberg's view of like what's to come in AI. That it's not necessarily going to replace lots of human workers. Kind of like the internet came along and social media managers came out of nowhere. There are so many jobs that are today are too difficult to solve or not worth your time solving that I think AI opens up to us. So just like, as you've worked at a tech company for the last 10 years, you never actually have less work to do. Like, only like having access to tools means that you have more work to go do. I think I don't subscribe to the, the trade after all that of like AI will replace people. I think for— there's a really good quote by Ross Simmons, who's well known in like the SEO space. And in short, he says that AI won't like replace, it should empower. So mediocre marketers, they're definitely a threat if they don't get ahead. But good marketers will use AI to go do work faster, great marketers will use AI to really like automate themselves out of like the mundane jobs. So I think there's definitely tiers to it, and like you need to be using the tools to at least keep pace, and then if you can really master them, you can start gaining, um, more like agency in your work. I think the other thing to ground the conversation in, and it was really helpful being in front of like GTM engineers through And the webinar Jared hosted during the week, but there's a lot of anxiety. Like if you go on LinkedIn, everyone is saying that they're using ClaudeBot or MoogBot, whatever it's called at the second to like do everything in their lives. Like I am very dubious that that is the case. And then even within a growth or marketing realm, people are like, I have this agent that orchestrates 10 other agents and they all just like mean my work is done in like half an hour. And it's kind of like that book, The 4-Hour Workweek. If you read it, it's not really 4 hours. So I think there's just a lot of BS out there. And it's just important that you're, you're testing, you're learning, you're experimenting at this moment in time. And I think then you're just like getting the muscle in, like, who knows where we'll be in like 6 months with respect to the tools and technologies that are available to us. So I just think if you're actively learning today, you're probably in the top 10%. And that's just where you want to make sure that you stay.

Yeah, I agree with a lot of that, Owen. And, you know, I just found for myself, it unblocks. So if I was trying to do a job before and I got to the limit of my knowledge, maybe that's where that project would stop and I'd move on to something else. But it just opens up so many more possibilities. And that unfortunately or unfortunately has meant I'm working more. And I think, you know, I've spoken with a lot of people. I think a lot of us feel like we're working more because we can. But I wanna double-click on the kind of last thing you said and, and really go deeper into separating that hype from reality, especially in the, in the, in AI-powered content. There's a lot of talk about like, hey, now, you know, I could just publish 1,000 pages now and let AI do all the work. So like, where, from everything you're seeing with your customers, where is content engineering? Where is that approach like genuinely a step change? And where are teams kind of fooling themselves?

Yeah, great question. It's probably the one I get the most. And it's, it's so interesting that it comes from people who are just dipping their toes in the water. And it comes from really big companies of like, where's the upper limit on like, how much I can let the AI take over? And, but if we take one step back of like what the intention of creating content is, and I think Eli Schwartz in his book Product-Led SEO makes some really good points and it's very like sentient. It's written 10 years ago before a lot of the different real estate even opened up in Google. But he makes the point of the search engines, their full intent is to surface the best context and content to the user. So if we bring that to today, whether it's ChatGPT or Perplexity or Google and all of their latest features, that's still what they're trying to do. They're trying to find for your given question or prompt, what's the best piece of content to go surface for it? And like what that needs to be grounded in and what we're seeing success from like our customers and our own content and what you can do today is like making sure that you're providing information gain to the LLM. So What does that mean? That means that you're giving it context that like the LLM isn't trained on. Like if you can imagine OpenAI is going to scrape all of the web pages out there every couple of months, like that becomes like what the average marketer or what the average prompt will return. What the search engines reward today is like unique content. And then whether you're on my side of the world, which is like content, or you're like a more of a deep in the weeds GTM engineer on like sales, you want to make sure that you're bubbling up the right context. And that's going to be totally grounded in like what your audience cares about. So we can go deeper into like what data mining is, where to find really good information. But that is what you want to ground all of your work in. And then you can start to use well-constructed agents on top of it. To figure out, great, like, here's some questions, they're topically relevant, let's create some pages. And but we should always— this is across any function— maintain your human in the loop. And that's where you will get like the 10 out of 10 quality rather than something like a 6 or a 7. So you definitely— people are listening to this— definitely do not go and just let your Claude bot or any agent go wild creating hundreds of pages for you. It's actually gonna do you more harm than good. You wanna make sure that it's grounded in your unique insights, and then do maintain like a human in the loop because there's gonna be pieces that your agent, until you develop it further, won't quite get right. And then that's where you're actually gonna spend more time improving the agent rather than getting it from like 0 to 80%.

If you think about this just explosion in how much content can be produced, because not everybody's going to listen to that advice. Like, there's gonna be like 10x more content out there, and yet I think we're not going to do 10x more reading, right? We're moving towards TikTok and Instagram and we're doing less reading, but the systems are cranking out more content. So, you know, certainly it's going to be the better content gets surfaced, but I wonder if kind of an inverse function happening here, which is that overall your metrics are gonna go down because there's so much more content, if you know what I mean. It's a much more competitive space.

Yeah. Yes and no. Like if we think of the SEO playbook from what, like 5 years ago, if you were creating SEO optimized content, the way to do it was maybe you have a page, you maybe your page ranks position 8, you would look at, hey, what's above it? 'What do they do right? Can I copy it? And then can I shove it into my article?' So it was very like undifferentiated and like you would constantly like hop between the top few positions. Now, if you kind of think of— there's this term of like cosine similarity. It's like how closely related is your piece of content and all of its context to the question someone answers. If you just continue down that undifferentiated path, everything is kind of in the same bucket and it all looks the same grayish color. What the search algorithms actually reward is what is relevant, but what is like standing on the top. I don't know how to continue the grayish metaphor, but what's like bright white that is showing like unique information. And just like us, like I enjoy reading both of your guys' posts on like LinkedIn because it's differentiated, it's like grounded in truth, usually like have some stats in it. And like just as much as an agent is trying to find that unique insight and like data points to go surface, that's what humans want as well. So I think creating great content for the like AI search era isn't all that different from creating great content that we all want to spend an hour or two on our limited time in the evenings or at the weekend actually consuming as well.

So you mentioned, uh, the latest hype cycles around cloud code and I don't know, do you see work moving from other tools to code it yourself, vibe code everything, or how do you think about purpose-built platforms? Are they gonna be winning in certain spaces versus the just vibe coded approach? Like when, what breaks when somebody tries to duct tape workflows together? Versus using, uh, using something that's where the system has been designed for it?

Yeah, great question. Like, two things come to mind that I have like personal experience, experience on here is like build versus buy, and then like bundling and unbundling. Um, like many times in the past, I definitely should have bought rather than build, uh, and I think the barrier to entry and barrier to entry on building has definitely come down, but also like the quality with which you have built, which you can build, has gone like way up. And then there's always this notion of like bundling and unbundling, which is if we take TV subscriptions, we all had like cable years ago, then it went to like unbundling, you have Netflix and Hulu. And now it's like, hell, like if I had all of these subscriptions, it actually cost me twice as much as if I just have cable. So you're always going to jump between in different industries, like almost purpose-built tooling and more like generalized ones. That's like the bundling, unbundling. So I think between those two domains of build it yourself versus buy and bundling and unbundling, we're probably in that transition of like moving to bundling together. And like you said, Jared, it's not fun having to stitch a bunch of tools together. And so I think we'll probably start to see like some tool consolidation. And what you'll keep in your tech stack are the tools that both serve the generalized automation purpose of like helping you stitch stuff, um, and also the ones that are like dedicated to a particular task. And what, what you can do then with that like smaller tech stack is have your MCPs or cloud code on top of it that really help you orchestrate the level below it and help you get the work done. So what do I mean by that? Talking to like the Ramp team last week, they just move fast in every direction. But what they're doing is they don't need all the tools that they have. So they're trying to say like, what are the ones that we have to keep that do particular job really well that we don't want to spend time actually trying to recreate? So like AirOps is one of those. They have like project management tools, they have in-house like databases as well. So they're now starting to create a layer on top of it that is like customized to what they need to go do on their like content team. And they use like MCP's Cloud Code to like orchestrate that level. And then that's where they're starting to like invest time. It's can they create this kind of like agent layer above what they've already invested in that just helps them to move quicker? So all in all, I think people are starting to push the boundaries and like what they can build themselves, but some things are like still a ways away from being recreated. Like, I know Salesforce and other, um, CRMs took a big hit, and that's definitely on a 6 or 10-year time horizon, um, more so than 20. So like, will a CRM be around in 20 years' time? We don't know. But will it be around in 6 years' time? Yes. It's like so hard to replace some of that stuff. So I think for the the easier jobs to be done, the stuff that isn't like super deep in your infrastructure, um, we will see a lot more people build custom solutions. But I think for the, the hardest tasks where like you have a tool in your tech stack that just does it really well, um, I don't see people going to like reinvent the wheel there in the next like 12 to 24 months.

Yeah. Yeah, even, uh, for those of us that have, uh, vibe-coded applications out there where people are using them, you know, it's, uh, it's still ongoing, uh, tweaking and maintenance, uh, uh, API, uh, changes, things go on and, and now, you know, you're suddenly, you juggle between your normal job and this applica— you, you know, you now find yourself a, a, a product owner or an engineer for these, for these applications that turned to perp. But are now kind of weighing you down in a way.

And I think that's, um, that's— people often jump ahead and they can make a lot of progress very quick, which is great. And that's what people should be focusing on right now. Like, am I learning? Am I making some progress? Like, the maintenance part is not fun. Like, it's great to just be able to, like, log in on your weekday morning and a tool works. It's not fun when I know from my days of building stuff in-house, you got a ping late at night and it's like, this is broken. Like our dashboards aren't going to work the next day. You need to go fix it. I think the, like how easy it is to manage those tools and softwares will become easier in time. Just like everyone is vibe coding stuff right now. No one was doing that, or a lot more people were not doing that 12 months ago. But I also think pick your battles with what you invest in. And what I mean by that is at my last company, which was like 3 years ago, as soon as like ChatGPT came out and there was like some early image generation models, we probably overinvested on trying to generate a lot of our ad images with those models and it just wasn't there. And we probably like tried to overengineer the system to make it get there. And that was like the wrong battle to pick at that time. Now you have like Nano Banana, which you literally just need to give it a sample image and it's like, I want something that looks like this out the far side, and it can do it for you. So like, you just need to figure out what can I move fast on today and what can I learn that I literally can get results really quick rather than needing to put in like a ton, a ton of effort. Because in 6 months, someone probably smarter than you right now will have it solved, and then you can just like reap the benefits.

Now in your, the team that you run at, at AirOp, could you tell us a little bit about it? And then for you personally, uh, in managing this team, what are you optimizing for within the team? Is it innovation and flexibility or, or, or cost, or, or is it something else?

Yeah. Uh, so today I run like RevOps, uh, GTM engineering, which I'm excited to talk about in a second, and like core growth, and also like product growth. So if something has growth in its name, like that's, that's what I oversee at the moment. And it's like quite a wide remit. But what I prioritize across each of those teams, and in my own work, is high velocity experimentation. And like, we, we run a framework where we consistently have ideas, we log them, we score confidence on them, and then we always continue to take the top, the ones that have the potentially highest ROI as the work that we go work on. So sometimes you've launches and you need to pause on certain tasks, or the highest priority tasks may take a backseat, but we very much focus on what can we do from building out MVPs and larger projects that would start to deliver results and importantly learnings over time. And why we do it that way is we just learn faster than anyone else. And like, sure, you could commit to one big project in a quarter, but if that project doesn't actually like hit the mark, you could have done all the right work, but it's not going to generate the results. I'd much rather take 10 shots on goal, 5 of them hit, 2 of them are failings that you learn more about your audience or the market on, and then you develop from there. So kind of giving you like a concrete example, you might have seen in the last week, Ramp, uh, released like their internal sales tooling, and that's something that they've been working on for years. And like how they actually, uh, developed that was again taking many, many, many shots on goal so that before they ever committed to what are the 10 buttons we put on this screen and the UI. They had like experiments known of like what signals work for outreach, who are the personas they go after, like what's the time of day. So like there were so many experiments that had to go in to generate learnings so that they could actually codify and build infrastructure around it. So that's— I'm at a much— we're at a much earlier stage than RAP, obviously, but we're in the high-velocity experimentation phase. Where we need to adopt those. We need to capture the learnings and put them into practice before we can start to build massive infrastructure tooling. So that across the team, and there are some infrastructure tasks which you just need to do right all the time. That's how we try and prioritize everyone's work.

That's great. Yeah, I know we used to always use the ICE framework. Is that— you guys call it that too? The impact, uh, confidence, ease. More or less.

I don't have like an acronym for it, but there's like 7 or 8 different fields people need to populate on tasks before we kick off the work. So yeah, there's like ICE and there's a few more.

Well, I actually, you can't call it ICE today. This is bad connotation. We got to change the name of it.

True.

Fair enough. All right. Switching gears. I'm excited to hear your thoughts on what AirOps is famous about, which is AI search, right? First of all, has the, in your point of view, has the market hovered or decided on an acronym for the new category, or are we still, is it still a debate?

No, we have not decided on an acronym. People prefer different things. I honestly couldn't tell you the difference between many of them. There's like, new ones like GSO and then like LLMO. Like what we just opt to use is like AI search or AEO because they're just like understandable without needing to get into like too much nuance. But I think in like 12 months, 24 months, we're just going to go back to calling it like organic search.

Love it. Yeah. Wondering why exactly platforms like G2 or Capterra haven't yet pick to also like a category, which is like still SEO, right? For now.

Yeah, I think it's just, there's so many people that have their like stakes in the ground for certain reasons in one area over another. And like, people are also just trying to like SEO-ify the game, as in like, if they come up with an acronym, can they create enough content around it where they then rank for it and they're trying to popularize it? I think that's like a very short-term game. Like people at the end of the day are trying to get their brand and their content in front of their audience, which basically comes back to organic search.

Absolutely. Yeah. People searching for people having problems, searching for solutions. That's the evergreen first principles. Love that. And again, from your front row seat, you know, like you're talking to the cutting-edge companies that are embracing, you know, like AI search and are building on AROPS. What have you seen being the biggest misconceptions from, you know, top operators to, you know, when approaching these issues?

Good question. Probably two things immediately come to mind. One is companies think that like AI content is bad or that they shouldn't do it. And then number 2, and it's kind of a dependency and it's related, people are like, we actually don't have something interesting to say, like we don't have data. And I think that's just maybe by the nature of like content teams not having access to it in the past. But going back to the first, it relates to, I think, like your question earlier, Jared, of like, where does the needle exist on like how much you can let AI write your content? People are very worried about that. I think it's like underlying anxiety of AI is there to replace my job. But if you think about most content, there's a lot of research involved, and there's a lot of like curation, and there's a lot of like writing, refining, editing, reviews, approvals. There's many different steps, and like, you're not going to enjoy all of it. And when we talk about like AI content, like all of that kind of gets like bucketed in together. But like, if I'm writing an article or like a memo for my internal team, Do I want to go do all the research? No, but I want to like send an agent off to like do 10 things for me and come back with the 10 best resources like that it can find in 60 seconds. It would take me multiple hours. So there's different phases. And I think the misconception is people think it's like zero or one. They have to be writing it themselves or they let the AI, quote unquote, like just do it for them. And The best solution is like a hybrid where it's how can you maximize the value of all of the intelligent people on your team throughout the content process. And like, there's some people that you want to reduce their involvement, and some people you want to like increase. So the reduction would be if you're Carta, and you need to have a legal review person at the end of every piece of content that you want, like that's not really probably fun for the legal person. It's definitely not fun for the content team because it slows it down. So like how we're seeing a team like that build an agent is they're working really hard with their review team. So in this case, legal, um, early on in developing the agents that they can codify what they care about. And then over time, as they get more and more reps in, the legal team are then starting to see content and they're like, I actually have no feedback. Like, this is like accurate, like anything I care about is like gut-checked and it's done before it comes to my desk. We're not saying then that they're just like pushing it to production, but like their time spent on review is like way down. So that would be like the reduction part to kind of solve the misconception. And then what gets increased is like access to subject matter experts. Uh, the hardest thing you can often do if you're dealing with like a technical product or super technical co-founders, actually get the information out of their heads that you want to communicate to a like-minded audience. And so like what we have done internally is a lot of our workflows are there to basically help us identify in-house who is working on this new feature, who's working in this domain or is knowledgeable. And that could be a salesperson, could be an engineer, And then we can then go and get that person to record like a 5-minute transcript of themselves to like document their subject matter expertise, and then that enriches our context. So bubbling all that together, and the misconception is that like it needs to be a 0 or 1 on letting AI do the work. It's not that, like the hybrid situation works best and Yeah, I think, I think that's what a lot of teams forget about. And then on the second side of it, it's like information gain. Lots of companies or lots of content teams think like, I see that this like information gain topic that you're speaking to is working. Like I need to differentiate my content versus what else is ranking or getting cited. And like what they've found hard in the past is accessing the data and it lives in, uh, like in-house in your experts' heads. It lives in support tickets. It lives in like your chat interface. It lives in sales conversations or CX conversations. It also lives offsite in people also ask questions or like subreddits. So with all the tools that are now available, I think it into Claude Code or ChatGPT, you just connect a lot of these in via like MCP and actually via natural language interact with the context. But the misconception is that like businesses don't have anything unique to say, like you actually don't exist unless you have a unique take. So I think a lot of the context that previously was very hard to access is now just like a couple button clicks away in Claude, or people use AirOps, just add in your integrations, and then that's what is your most valuable information to turn into content. So definitely misconception around zero to one on how much you let the AI go do the work. And so hybrid is the best situation there. And then people think that I don't have anything interesting to say. If you're talking to customers, if you're in the marketplace, you always have something interesting to say.

I love that. I mean, I'm super bought in on the modularity or the fact that AI to human relationship is a switch, you know, it's not, it's not a switch. It's a, you know, like you can actually like modulate it and you should be depending on which workflows you're working on, you know, depending which outcomes you're working towards too. And that's actually the mastery where also like experimentation, you know, like, or the learning from experimentations comes in, right? You know, like you, you learn how to modulate, you know, like these levers based on the learnings that you get as you put the reps in. Right. And I think like that's a super fascinating, a super imperfect science as well. You know, like it's more of like an art, if you will. And you build muscle, you build taste, right? You know, with, with experience, you know what good and great looks like, you know, how to discern from bad and less good, right? So food for thought there for sure.

Totally. And the example that stands out, and I was laughing earlier when I said I wanted to get into like GTM engineering, I was putting together our onboarding doc for our new GTM engineer who starts on Monday, who I'm super excited about. And like I was trying to— there's so many things that we need to get done. We also need to be realistic of like the path that we're on. I think that that's where our framework of like high velocity experimentation comes into play. Because like right now, when this person joins, when Ben comes in Monday, we're trying to make sure that he can compound learnings. So there's like, let's, let's like help them understand what we have done so far. And then how can we set them up for success such that when we look back in like 90 days, like in his head exists a bunch of learnings that then we can start to like codify. So I think when I work with my team, oftentimes it's very like easy and quick to say, let's go automate this. If you actually don't know what works and what doesn't, you haven't earned the right to go automate it yet. I think like that needs to be earned again, we've done a crap ton of webinars, but the first 5 or 10— I think you were our first one, Matteo, actually— and our first like 5 or 10 were super manual. Like, I was up late trying to like write the content. But then as you get the reps in, you understand what great looks like and what hooks work, what format works, and then you like can actually go to codify it. But, um, very much you need to make sure that you're maximizing for learnings before you're committing to building out like a Claude project or anything even like more simple.

And I love also, like, I would love to double-click a little bit into, you know, like you talked about how all this context that makes your opinions distinctive and differentiated and your ultimate content winning, so to speak, live in, it's very unstructured. Lives in people's mind or in siloed places like Gong or Slack, you know, like, and, or, you know, customer conversations. How are you guys trying to have, you know, like set up some listening tooling or ongoing input feedback loops to actually like aggregate and structure these, this win, this, you know, secret ingredients, let's call it that way, you know, like, and then funnel it back into your content workflows, whether that's, you know, like your codified learnings from your new GTM engineer, like week by week, or whether that's, you know, like, why people love Airops from your customers?

Good question. I'll break it down into two things. And I always find it really difficult consuming some information on LinkedIn where it's like, hey, I've built the thing I spent 6 months working on, here it is, like, good luck recreating it. Like, I like the, what can I do this week sort of answer. And then I can get into like the longer form answers. Like, what you can do this week is just try your best to mix some qualitative with some quantitative data. So example of that is take your sales team's last 10 call transcripts or take your last 50 support tickets from Intercom. You probably have access to Claude or ChatGPT. Throw them in there, ask it to find trends, topics, insights, and then format it for you as questions. And then again, like you need to have some like qualitative, quantitative data in there as well. So it's great knowing like questions, but if you're going after search in particular, you wanna make sure that there is like volume there too. I will say that if you're trying to find an in with a prospect via an email, there's going to be a different way to go and approach that topic. But taking your GSC or GA4 data, adding it back into Claude or ChatGPT, asking it to interpret, okay, from what people— what, what— from what's coming up in conversations and what you know about like where the potential volume is, where are my opportunities? That is like what you could do this week to figure out what are the 5 are 10 pieces of content you can go work on. If you want to take that one step further, integrate your Slack and then also say, hey, now that I've prioritized the 5 or 10 pieces, now find the key things that come up internally at the company around those topics. And then, you know, actually Matteo Tittarelli was talking about this like MCP feature we're releasing. Let's go ask him 5 questions and that'll improve that article. So very quickly you can go from qualitative data in a tool, some quantitative data in like GSC, GA4, if you have a different tool, by all means. And then like take some internal expertise. That is like your, your one-week project. How you start to build on that over time is kind of like what I said earlier, you earn the right to automate wherever you start to see value. Maybe it's, oh, like this piece of content performed really well because I maximized for query volume, or this works really well because it's topically relevant to someplace that we have domain expertise. And then you can start to figure out, great, let me create a really good workflow that lets me stitch the last 100 sales conversations from Gong into an Airops or a similar tool. Then I want to enrich it with my AEO or SEO or AI search data, whatever the acronym is. And to figure out how do I prioritize the most recent questions that came up. And then from there, again, you want to figure out like, does refresh work for my business right now? Or should I be creating net new content? So going back to the methodology we live by, it's very much, can you start to focus on aggregating learnings and then building your automation bit by bit after that?

Beautiful. Thanks so much for sharing.

You're in such a unique position, Owen, because you get to work with some of the best companies out there and see how their teams have been evolving, how, what their, the mistakes and the, the lessons that they're learning about hiring. So I bet from that, AirOps has been able to make adjustments on your end. So you talked about your team before. Is that everything that has growth in it. So you mentioned RevOps, go-to-market engineering, now core growth, product growth. I don't think you mentioned content engineering. So I'm interested to hear sort of what are— where does content engineering live, or content creation live at AirOps? Uh, well, let's start there and then I, then I'll work my way backwards.

Sure. Uh, so content engineering lives under like our core growth team, and like how it might evolve our company over time will change. So again, if you think of a, a bigger company like Webflow, they'd like a dedicated SEO team, and Vivian is like their lead. She became a content engineer. If you look at another like fast-growing org, like a, a Ramp, they've been able to like hire multiple content engineers over the last 12 months. And like how they're constructing an org is quite different. We're still growing and we've invested in like content engineering as like it's a standalone function from day one. It's supported by people who have deep knowledge in SEO and deep knowledge in like our audience and our market. So like product marketing and like SEO backgrounds. And I think that's, that's how we will continue to grow it. We need to make sure that our content engineer is supported with the, like the governance data. So like from Matteo Tittarelli's question, like, are we able to pull in all the right context? Can we enrich it with our tone of voice and like our brand persona, which is going to come from like the brand team? And then the content engineer is like really helping make the magic happen and creating the content. And why I think, well, for us anyway, it sits under growth is that it's a, it's a growth lever for us. We're creating surface area for our brand to get found and for people to, to find us through like search engines. So it is a huge growth lever for us and like that's why it sits under growth. As we grow and we have like more defined like marketing functions, it's probably going to evolve into something like what a Webflow or Ramp does. But for each different org size, you're going to have a different kind of like fit or need.

And for that team, what is the true north metric? Is it just visits from the content? Is it specific keywords you're going after? Traditional? SEO type of metrics where you want to rank for them? And now in the LLMs, is it number of assets published?

Yeah.

And what— how do you, how do you measure it?

Yeah, great question. This— I do like roundtables every once in a while with some of our like customers. It's the number one thing that comes up. And then the follow-on question is how do you attribute it? Definitely at the top of the funnel, what we care about is like mention and citation rate. So think of it as like share of voice. There are certain questions that we care about, and when we look at the names that are generated in the responses to those questions, are we getting mentioned? Like literally, are we in the conversation? That was like the number one thing that we care about. And if we're appearing in more conversations, how we then look to attribute it is through our brand search volume going up. And if more people are literally like searching for AirOps, sure, there's underlying lots of brand activity that we go do, and there's lots of other channels that contribute to it. But as we've seen direct kind of traffic come in— sorry, as we've seen less traffic come in from Google per se, we've seen our branded search traffic go up. And then I think it's like The correlation of the two of those just show us that AI search and what we're investing in content engineering is working. So TL;DR is kind of share of voice percentage for the queries we care about. That's one big north star metric. And then our branded search queries, that just shows that we're becoming more and more seen by the people we want to. And are you able to use AirOps tooling for most of of that analysis? For all of it, which is the nice part. So, uh, I never like to use the word dogfooding. Uh, internally I'm known as the guy who drinks our own champagne. So drinking your own champagne is my preferred method. Um, yeah, so we use AirOps to like fully orchestrate everything from defining what questions we want to go after, tracking those from both an SEO and AI search perspective. That helps us then prioritize what content we create. We can create the content or refresh the content in Airoops. We can publish it and then we can do the measurement to complete the full loop. So end to end, we're using our own tool, which is great. And like across the full intended spectrum.

And is that something that our audience can go find your recipe for? Like, have you published that? Have you done some content on how you drink your own champagne?

Yeah. We have, um, we have a series in our blog on like how AirOps uses AirOps. And I think for a GTM engineering audience, we have like 2 or 3 of how we've, um, put together some like deep research and then competitive, uh, research like workflows, how we do some LinkedIn and social posts. And then for what I mentioned there, content creation or refresh We have many more that were covered, um, by us in our own articles and then also on like webinars with some like amazing customers. Awesome.

Great, uh, great resources. I think I've seen some of it, but I'm going to have to go back and find the rest that I missed. And then now back to the go-to-market engineer that you're hiring, how are you envisioning that role as separate from whatever you have today in terms of RevOps and even from the core growth? Like where are the, where are you going to draw the lines?

What are they? Really going to be responsible for? Yeah, just in where they— just in terms of where they sit, like organizationally, I want them very much focused between both the sales and marketing team. So why is that? Like, they should have deep contextual awareness of all of those things because top of funnel, they need to know what brand growth and marketing are seeing working. Middle of funnel, I want them to sit literally like next to the SDRs who are on the front line. They're getting rejections, they're getting hopefully demos booked every single week, and then like working really closely with like AEs and our leadership to figure out like what's actually driving revenue. So in terms of, I need them to have like a 360 view. And right now the best place where they can get the most of that degrees of freedom is within core growth. And what I'm very much focused on having, like Ben, um, like said, is his goals are around almost developing like what Ramp has, which is like the internal tooling that makes the rest of the team super efficient. So there's like a couple of different aspects that underline that, which is like, do we have the core infrastructure? And again, that goes back to the earlier question of like, what can you vibe code or what should you not? So like, there are certain tools there that that we've built in-house that will continue to invest in. There's other stuff like Clay that will just get out of the box, like why go reinvent Clay right now? And so that's one big thing. And then the second goal is around increasing like SDR or general sales team efficiency. There's lots of plays that we've run that we should just at this point go and automate because we actually don't need a human to go run them. There's other places where we still need to test further and experiment to increase our learnings to go later on, automate. So right now it's very much like shore up and make sure our infrastructure is scalable. And then number 2, can we make the existing team more efficient? And there's like multiple different vectors to that.

This is really interesting. And so for Ben, he's gonna love this podcast because we're talking about him. It, it's more the success metric for the, because this is gonna be of interest to the go-to-market engineering audience, like what they're, what managers are, are, are evaluating them on, expecting them to deliver on. So it, you, you mentioned it's more about building the systems that sales and marketing can use, not so much putting a pipeline number associated with the GTM engineer's activities, right? It's more about enabling everybody else.

Yeah, I will say that we do care about like pipeline generation or like demos booked. I think it's good to be always grounded in that as a number because otherwise, kind of like for some content teams today are still optimizing for traffic, albeit the traffic might not convert. You wanna make sure that there's like some goal towards like a number associated with revenue, but then we really wanna make sure that we're using our human man-hours as efficiently as possible. And that's like very much what we're, we're trying to get better measurement on and then, um, have like Ben goals towards as well. Like if we're hitting the two things, if our team is becoming more efficient and the revenue number is going up, that's the win-win. Then we can start to like move levers in different directions. I love it.

Well, we're gonna have to have Ben on in season 3.

We'll make it happen. Yeah, we'll tell a story of just a great 6 or 12 months of GTM engineering and AROPS.

Yeah, do follow the journey. Well, I think we're coming to the end of our allotted time here, but this was a great conversation. Um, gave us a front row look at how the shift is occurring from SEO to AI search. And what is really working in practice? The teams that, that, that win, it doesn't sound like they're the ones necessarily just publishing the most, but the ones that are grounding the AI in proprietary context. And it also sounds like companies are investing more in building their own internal database, pumping things in from these tools. So that proprietary context is key. And the other thing I recall is mentioned a few times is, is a strong focus on keeping the human in the loop. So let's say Owen also showed us that the gap is widening between the average operators and those who are aggressively experimenting as your team is doing, building your custom workflows and then only automating once they've truly earned the right to automate by knowing what works. Owen, this is great. Any— how can our audience get in touch and find out what's going on in your life and in the progress of my life.

I don't know about in your life, but in your work life, I joke. Um, if people want to learn more, definitely go visit, uh, aeroops.com. We, we do lots of webinars, and we, we love just like educating the market on exactly what I covered, uh, throughout this episode. And then yeah, we're constantly hiring. We'll always continue to look, to look for amazing growth and then GTM engineering talent. So come find me or any of the team on LinkedIn more than happy to chat. And we also have, um, like free trial, free access to the platform. You get everything for essentially as long as you want. So, uh, go sign up for an Airps account. If you want any help, come DM me. I'm easily findable on our website or on LinkedIn. And then my main call to action for, uh, this audience, given it's a technical one, try out our MCP. We're actually going to do some pretty cool giveaways over the next 30 and 60 days, uh, when people generate amazing, amazing comes and use it in interesting ways. So try out our MCP, happy to add you to like our beta group. And yeah, excited to see what people build.

Go drink Owen's champagne, please. Yes.

All right, fantastic. Thanks, Owen. Thank you, guys.

See you later.