Latest / Found in AI: AI Search Visibility, SEO, & GEO / What Is a Content Engineer?

Transcript

What Is a Content Engineer? · Found in AI: AI Search Visibility, SEO, & GEO

• 0:41
  (Transcribed by TurboScribe.ai. Go Unlimited to remove this message.) Everyone's talking about AI replacing writers. Listen, I've heard those conversations, I have been a part of them, and for a hot minute, I was also very worried about this.
• 0:51
  But the smarter conversation is about the new roles AI is creating.
• 0:56
  Welcome back to Founding AI, the show about getting your brand found in AI search.
• 1:00
  I'm your host, Cassie Clark, content strategist and CMO.
• 1:05
  Today I'm taking a little bit of a different turn. I'm sharing a conversation with Josh Spilker, head of content and SEO at AirOps. We unpack one of the most overlooked roles emerging in content right now. That is the content engineer. If you've been wondering what it really looks like to build content for AI-first discovery, how to future-proof your own workflows, and what that role, the content engineer, looks like within a content team, this one's for you. Here's part of that conversation.
• 1:35
  Okay. First, my name is Josh Spilker. I live in New York City. I've been here about two years.
• 1:42
  I work at AirOps as the content and SEO lead, and I've been here about 10, 11 months, something like that. So AirOps is a content operations tool to help you build workflows for your content, but also to help you see how you're appearing in LLMs. So we have a visibility dashboard as well as workflows to help you take action on those insights that you find.
• 2:10
  So before you joined AirOps, what was your pulse on AI visibility or anything like that?
• 2:15
  Was that foreign, or were you already digging in?
• 2:19
  Yeah. Great question. So AirOps started as a workflow builder, specifically around automating workflows and things using LLMs. And so I've used AI for a little while to help create not only content, but briefs and drafts and things like that. So I remember using OpenAI Playground two or three years ago. I was using some other tools. I was experimenting with Cloud and Gemma, I think is the name of the presentation tool, things like that. So I had some familiarity, I would say, with some AI. And then I was also setting up revisions and drafts, or mostly helping me edit content using AI. So I'd just be like, hey, chat to PT, how would you change this? You're an expert at X, Y, and Z. What would you add to this? And so I was really using it as more as a stress test. And so then when I started at AirOps, I thought it was a good fit. And it just seemed to be cutting edge for what I saw with my background in content and SEO teams for how things were changing.
• 3:25
  Yeah. So you were kind of looking into the future ball and you saw AI. AI is coming.
• 3:33
  Yeah. I would say one of the leaders that really got me onto it, his name is Luke Thomas. I worked with him at Friday. He went on to Zapier and is now doing some new companies. But he was like, hey, we really need to look into this for some of its possibilities, like play around with it, see what kind of drafts and outputs it has. And so that's how I got introduced to OpenAI Playground. That was probably like 21, 22. Yeah. Yeah. So it's 2025 and it seems like it's changing again with agent. Can you release that like last week? Oh, yeah. The chat to PT agents.
• 4:05
  Yeah. Yeah. To help you kind of do tasks for you. Yeah. Yeah. I've not looked into that yet, but it's on the list. That seems helpful. I put helpful cautiously. Right. Right. I haven't used that quite yet. I don't know if it has access to now, but yeah, people in our office are definitely buzzing about it for sure. Yeah. It's all over X. So I wanted to talk to you about content engineers, because in LinkedIn feed, you're really the only one that has labeled it this way. One of your LinkedIn posts, you kind of mentioned that traditional content organizations were built for SEO and blog calendars, but not AI synthesis. So what does that shift look like in practice when we're thinking about optimizing for AI engines instead of actual search engines like Google? Yeah. I think there's two things happening here at the same time that SEO and content seems to be aware of. First off, it is just the capabilities of AI workflows, helping with like those manual processes. And so one thing that I always refer back to, and like my previous or in my roles as a SEO content person is like internal linking, for example. So I would have like this massive spreadsheet, all of my library URLs, where they needed to go and how they would connect. And so then I would map all of that out on a spreadsheet and then just like me and an outreach specialist or whomever else I could get to help start inserting those in. Because we all know that you need to update. We all know that internal linking is important. But of course, the more content you have and the bigger your library is, like the harder that becomes, even though we know it's like an SEO best practice. So one thing that you can do with AI workflows and AirOps, for example, is to like help identify where content needs to connect based on topics, based on keywords, based on like whatever guardrails you set.
• 6:03
  And then you can also create those anchor or identify that anchor text and then insert links and then like push it out through your CMS. And so that's an example where every content marketer or SEO has had to do that at some point. And it's using AI to do it. And it's not affecting your quality or your editorial standards. It's not writing. It's none of that. It's like a very simple process that we all know needs to be done that you can now do that accurately and on a large volume of content. Of course, you need to like double check and just make sure almost like a product mindset, a QA type mindset on that to make sure that the actual links are getting there and are in the right place. But you run a small set, 10 to 20, and then you expand it and you expand and experiment again. So that is just like one example of how I would say AI processes and workflows to really help a content team. And we can start thinking systematically about that.
• 7:00
  Then the second part, I would say, is what you brought up about peering and LLMs and tragedy and perplexity. And what we're finding is depending on your industry, but a lot of like kind of cutting edge industries really need to refresh their content quite regularly. And so then running workflows to help identify, but then also to help update that content is really important. So for example, one company we work with does a lot like in regulatory industry around like different states, different counties, all of this. And so it's kind of hard for them to keep up with like all of those changing regulations. And so you can use AI workflows to help surface those changes that are happening, and then to find places maybe in your current content to where you might need to update that. So there's a lot of like use cases. And so all that to say is keeping that content fresh and updated helps you then to appear and tragedy and perplexity. So it's kind of like a twofold thing there. It's helping you with the processes, but it's also helping you appear in LLMs regularly. Hey, quick break from this conversation. If you're enjoying this episode, make sure you hit subscribe so you don't miss the next one. And if you want to get insights like this straight in your inbox each week, sign up for my newsletter. It's in the show notes.
• 8:16
  Okay, back to it. Yeah. So one of the things I was talking with Nathaniel last week, we were talking about doing these content updates. And then we were also talking about how some of the models are trained on data sets from forever ago. I say forever ago, it's just, you know, six months, a month, a year or so ago. But it's also looking towards like Google and other things. So what was like the recommendation for how often you should be updating your blog posts, content, whatever?
• 8:46
  Yeah, so I've got a chart I can share. I'll try and find it real fast. But it's like, depending on your industry, it really depends on how often you need to update your content. So fast moving industries, we saw you need to update your content every like three to six months.
• 9:03
  And those industries include like SaaS, finance, news, the freshness window for other industries like real estate or ecommerce manufacturing, maybe like a little bit longer. And then things like travel, lifestyle and healthcare, we saw like six to nine months. So for instance, I like to use this example, like the treatment for diabetes is probably not going to change like super fast, like there are going to be advances, there are going to be changes. But generally, like the best practices for how to handle that are not going to be like rapidly changing like day by day to where it's like something like news obviously is changing pretty quick. And then you have SaaS companies that are always like kind of adding new features. That's the space that we might really deal with is like enterprise SaaS companies. They're like adding features, they're changing their pricing page, they're doing all of these things, they're putting out new research.
• 9:55
  So that is changing a lot faster. So I like to think of like that spectrum of like kind of identifying what industry you're in, and understanding how fast it changes like education, nonprofits, they don't change as fast either. So it's really just kind of judging which one you're in. So that said, yeah, the models get grounded at different times, they're going to have to reground themselves. I don't know if that's the actual word for it. But like, they're going to need some new information. And then more and more folks are hitting that like web search button on Chachapiti in particular, but Perplexity always, from what I know, always kind of has a live web web search. So they're going to be updating. And then of course, AI mode, AI overviews and Google are going to be like updating around those industry benchmarks as well.
• 10:43
  I wonder if they're using the same ones like Shirley, Shirley, like they're thinking about if you're in SAS is changing quite often. So it's like worked into like AI mode models for like Google or whatever. Surely. Yeah, I'm not 100% sure on that.
• 10:59
  Yeah, but that's probably like a good recommendation. So let's talk about like AI search in terms of like the buyer journey, and like single answer streams. How should a content engineer think about that and influence within those systems when the clicks are gone?
• 11:16
  Yeah, sure. So it kind of goes back to the example you brought up. It's like you have to continually refresh your content. And if you're having more content, because the longer tail of searches is widening, and there's the volume of searches is widening, then you're going to have to keep maintaining that content. And what we're seeing with AI search is that it is more conversational and is more specific as far as the queries go. So in the past, you might have just said like best project management software into Google. But now you're like, I run a 10 to 20 person marketing team, enterprise SAS, please help me figure out the best project management software with dependencies, you know, with all these x, y, z, with Gantt charts, whatever. So you can become like a lot more specific. And the research is showing that queries in LLMs are much longer than they are in Google. Maybe that'll shift as we kind of get used to LLMs in AI mode, maybe Google eventually those queries will be longer. But all that to say is like, your pages are coming up more often for these longer tail searches. So you might have to like update your pages a little bit more often to stay in there to address these like longer tail search queries.
• 12:26
  Because what we're also finding is like people are doing their research more intensity and perplexity before they ever hit your site. So then that top of the funnel research stage is being like stripped away. These LLMs are giving kind of that initial research phase. And so then when people do land on your site from these, they are more likely to convert. So we're seeing like smaller amount of traffic, but like higher conversion rates that come from that. And a really good example, we have a case study on this is Webflow has seen this happen quite a bit to where they're almost getting like, I believe it's like eight or 10% of their signups are coming from AI chatbots. And so the volume is much lower, but that's like a pretty high number that they're seeing. Yeah, that is a high number. So we've talked about bottom of funnel content quite regularly on this show. But I want to take a second and define actually what a content engineer is.
• 13:23
  Because I mean, that seems like it's still a new title, maybe. So when you talk about that role, do you see it as like closer to SEO ops function, data engineer or marketing technologist or somewhere in between all of that? Yeah, I think the marketing technologist is good. I define it as like a hybrid of strategist, technologist and editor who can build systems for quality and impact. So it really is all about a systems thinker. They can design in content systems for that. And they build workflows, they measure velocity and visibility. And they may even come up with like kind of their own little low code or no code tools, which you can do in AirOps or like custom GPTs, for example. It's not just about AirOps for this, it really is just like leveraging AI workflows from like whatever tools that you want to use. So content engineers are thinking in words, they're thinking in outcomes, but they're also thinking in systems that deliver both of those. So what kind of tech stack or skill set do you expect a content engineer to own?
• 14:31
  Are we talking like Python scripts, APIs or prompt engineering inside of those tools like AirOps?
• 14:37
  Yeah, we're finding like a mix of people coming to that. So I would say that they're going to own a few of those tools. It might be like, how do I phrase this? It kind of depends on the person, but also the organization. So I really see the person owning some like no code and low code abilities. Maybe they know some like Python or JSON, maybe they don't. It's just really about building that out. You don't necessarily need to know those things to build in AirOps, for example.
• 15:07
  So it's like a combination of just like the person's background, I would say, and what they're most comfortable with. But what we're seeing is like kind of this convergence.
• 15:14
  So we have like writers and editors who are now becoming content engineers. We have like more technical people who have built things in the past, also like kind of applying that to content.
• 15:25
  It just kind of like depends on that, I would say. We have people from all of those backgrounds.
• 15:30
  Yeah, so how does this role differ, like in plain terms, from like traditional content operations or even Redbox? Like how is it different?
• 15:41
  Yeah, it's really about systems thinking is how it's different. So it's thinking of content not as like one-off blog posts or even like campaigns, but like really thinking about that like system wide like updates that we need to make. And then doing things, maybe like creating more templates to get things done, almost creating like product requirement docs to like help build out your content. Like that's something I hadn't done until like recently. It's just like really treating that content as a product to get out the door. So I think that's a mindset shift for a lot of writers and editors. But then it's like, you know, we have brankets and knowledge bases to where you take some of your like call transcripts or research your past blog posts and we're like grounding your own model in that and grounding your own workflows in that. So it's like this kind of, our CEO here calls it a context librarian to where like they're maintaining this library of product information, of company information, of blog information, of research, of content, of point of view and thought leadership to then have them access. So we're really seeing a lot shifts here of like new roles for content leaders to have. But I mean, it's not only like new roles, it's like upskilling and kind of some shifting of roles as well. So that was going to be, that was going to be my next question because I know a lot of the listeners of this podcast are writers who are interested in or marketing managers who are interested in where AI search is going. And a lot of the things that we've been talking about, it's like the skills like off podcast, like the skills that we need to like stay ahead of this. So when we talk about that, what does that look like in practice? Are we talking like structured data, the content libraries, retrieval date layers, blog posts, and how does that all fit into like the skills that we need to stay fresh with all that?
• 17:37
  Yeah, it's probably a combination of all those. It's kind of a cop-out answer in some ways, but I mean, what we're seeing is that original data still matters. So doing that research, like having that subject matter expertise layer is still really important. Point of view is really important. So you still need writers and strategists to determine that because you're trying to like cut through the noise even more because like, look, the models, like we said, have taken out like top of funnel in a lot of ways, which is affecting maybe some like traditional content writers. So now it's like, what does that value add that you can use to like cut through the noise?
• 18:13
  And then it really is like that development of the middle funnel and bottom funnel content as well. That is like really top-notch. So you might be creating more of that, but you need to be like more specific on it. And so you can use some of these AI systems like we've been talking about to help with like your research, your outlines, your initial drafts, and then your actual writing may look exactly the same. And that's what I try to chat and talk to people a lot about as well, is it's like, what are some things that you can automate and what are some things that you shouldn't? Everyone's going to like kind of make that decision themselves, but you may automate only some parts of the process, but not all of it. It's not that AI is like an easy button that you push and then everything is done for you. There's still so many guardrails that you have to put in place. And so then on the back end, like the repurposing aspect, AI can really help you along with that as far as like summaries, TLDR, maybe like taking out some key bullets. I use a workflow to help me identify good quotes from like talks like this in the webinars I do. So it's not like one thing that AI is completely replacing. It's like replacing some parts of the process or maybe helping you in some parts of the process. So then you can like grow your output and to like do all these other things that you've been maybe wanting to do on your back burner list, but now you can maybe kind of finally get to it. So if you are building out a small team, like marketing side, content marketing side, would you hire someone to just do this part with the content engineering part? Or would you have that with content manager doing the same thing, like juggling that within their role? Yeah, we are actively like kind of discussing that on our end too. Like we do have a content engineer who helps us do this, but like I know how to use some of the workflows. Some people are better at it than others, like building them. So I've built some workflows. My content engineer or our content engineer, Ocean, like really kind of takes it to the next level. So it really is kind of that like mix and it depends on the board. But yeah, I mean, I think everybody needs to consider how they're doing this, especially if like search is a main channel for you, but also just like the repurposing angle, like I said. So finding a contractor to help you do it, whether it's an AirOps or another tool, I think is super important.
• 20:38
  Understanding some of these skills is important just so you know what to hire for or like what to do. So it's like, yeah, I'm going to get an AirOps. I'm going to get in cloud code and like mess around. So then I know like kind of what I'm looking for and like what some of the capabilities are and then go out and hire for that. But yeah, I mean, this isn't like going away. And so I feel like equipping teams and to consider this as a role is very important, whether it's like a blend of upskilling or having a dedicated person for it. So last question, I know that you made a post recently, and I don't know how recent this was, because all my days are running together about AirOps University. So if someone listens to this episode and they're thinking, gosh, I really need to like somehow get those content engineer skills, is AirOps University a place they can go to start learning how to do some of this? Yes, for sure. AirOps University is free.
• 21:32
  It's like hosted on Circle. We have all of our live cohorts that we've done recently. We kind of put those into a self-directed course that anybody could take. You can also set up for a free AirOps account to like start building with it. And then we have examples and folks from companies like Wiz and Carta and LegalZoom who have already gone through our cohort. And I think some of their work, we're going to be adding maybe some of their videos and things in there. I'm not exactly sure on that, but like, yeah, we have all that course content to help you walk through it step by step, how to map out the processes or just like parts of your process that can be automated first. And then the great thing about a tool like AirOps is like you can customize it so it's not just like out of the box. You can really like make it fit for your use case, connect it to your CMS, all of those things. If there's one thing to take away from this episode, it's this. AI is not replacing content teams. It's rewarding the ones who learn how to build systems that scale. So if you're a writer or a content manager and you're looking to upscale, start small. Pick one part of your workflow, whether that's briefs, outlines, updates, then build a repeatable AI process around it. You'll start thinking less like a writer and more like a content engineer. I'm Cassie Clark and this is Found in AI. If you want help future proofing your content strategy, you know where to find me. Hop over to LinkedIn or find me at CassieClarkMarketing.com. Remember, the future of content isn't who can write the most.
• 23:08
  It's who can build the smartest systems to make every word work harder. I'll see you next week.