In this conversation, Noah, Eoin, and Josh discuss the recent launch of AirOps 2.0, the challenges and successes of collaborating with industry leaders like Kevin Indig, and the platform's focus on helping marketing teams optimize their content operations. They delve into the importance of grounding data for effective content creation, the evolving role of content engineers, and the distinction between agentic and hybrid workflows in content production. The discussion highlights the significance of human-AI collaboration in achieving high-quality content outputs. In this conversation, Eoin Clancy, Noah, and Josh Spilker discuss the integration of AI in content creation, the importance of effective workflows, and the role of professional services in enhancing user experience. They explore the challenges and opportunities in startup growth, the significance of API integrations, and the strategies for delivering value to customers while navigating pricing complexities. The discussion emphasizes the evolving landscape of content engineering and the need for continuous adaptation in a fast-paced market.

Chapters

• 00:00 Introduction to AirOps and Launch Week
• 02:58 Collaboration with Kevin Indig and Product Development
• 06:09 Understanding AirOps: Target Audience and Features
• 08:51 Building Relationships with Industry Leaders
• 12:07 Differentiating from AI Slop and Programmatic Content
• 15:02 Grounding Data and Content Creation Workflows
• 17:48 The Role of Content Engineering in Modern Workflows
• 21:03 Human-AI Collaboration in Content Production
• 23:59 Agentic vs. Hybrid Workflows in Content Operations
• 30:11 Designing Effective Workflows
• 32:25 AI in Content Creation and Review
• 35:34 Leveraging Data for SEO Insights
• 39:29 Building a Professional Services Team
• 45:15 Startup Growth and Market Traction
• 49:21 Integrating APIs for Enhanced Functionality
• 53:58 Support and Customer Experience
• 55:03 Pricing Strategies and Value Delivery

Campfire Chat Transcript

Noah: edition of the Campfire Chat series. I'm really, really lucky today to be joined by Eoin and Josh from AirOps. Guys, it's amazing to spend some time with you. You've been busy.

Eoin Clancy: It's been a busy launch week. went live with 2.0 on Monday and it feels like we've lived a month since.

Josh Spilker: Yeah, it's

been great to be a part of it.

Noah: Tell me a little bit about what it's like working with someone like Kevin Indig, who does a double post where he posts something on growth memo, which I love that article. And then he posts another video with direct links into a spreadsheet and a link to AirOps to run something. Did that stress any systems at all?

Eoin Clancy: Yeah, stressed a of systems and people across the org. So it was really fun working with Kevin. And I think when we were originally on what, you know, a pre-Christmas, we were just getting some of our different motions going. And it's really nice now where we have some people in the network who we wanted to work with. And now they're asking to work with us a lot more, which has been a nice changing of the tide. But yeah, it was really interesting yesterday. We...

prepped with Kevin for maybe two weeks. had a concept that he wanted to realize. He had a good sketch in a workflow in AirOps. And then we helped him get that to the finish line.

So there was a lot of work from one person on our team in particular, Nicole, who's on our sales team, but she's probably one of the best builders at the company. So we got that live for end of day Monday. Kevin wrote his own article and it was his baby, his original concept.

So he pushed out his growth memo and then like his LinkedIn post. And within maybe the first 30 minutes, and it was just the beginning of the day here in the US, we were like thronged with requests. And they made like 500 individual URL submissions in the first 30 minutes, which was a lot more than we expected.

And behind the scenes, like where to a user, they put in a URL and out the far side, they get a very complex document.

that whole process requires money, money, money steps and like a framework on our side. So basically the day was spent running off to different vendors, trying to increase our rate limits, trying to basically remove all the bottlenecks. So the last 24 hours has been trying to work with Kevin so that his audience knows how to get a version of the workflow that will give them an output.

And now we've turned it into a big part of our product. So.

Maybe later when we show it. I think it just went live. I just got confirmation from our product person. It's in the dev instance, but basically we turned Kevin's idea into part of our product.

And now we need to let everyone know who basically tried and failed over the last 24 hours to get their output, but it's now available. So it was really quick, typical life of the startup. You put something out, does a hell of a lot better.

And then you try and like make a...

make wind follow the hills shines. So it was really good. Yeah.

Noah: Rad OK so AirOps what is it?

Eoin Clancy: Yeah. So originally started as a very horizontal platform to work with LLMs. We've really started to focus on, it's our core audience, is like content marketers and marketing teams. So at its most basic, we try and help marketing teams win the battle for attention across all of these channels that are emerging. So like some enterprises are heavily like socially focused.

Josh Spilker: You

Eoin Clancy: Other ones that we work with like SEO and organic traffic is like what keeps their growth team up. So what we focus on is facilitating those teams to build the workflows and the content operations that power their teams. And just adding onto that, we launched this week where a lot of our messaging is towards what we call AEO or answer engine optimization. Everyone here will be well familiar with the HubSpot story.

We hear a lot of customers coming into us where they've had massive impact to their traffic from AI overviews. So people are trying to figure out now, how do I write content that will appear in the existing AI LLM searches, like perplexity or AI overviews, but also where they have existing content. And that kind of lends into what Kevin's original idea was, where they have existing content.

How can they tweak it? How can they refresh it or optimize it?

so that it is cited in all of these different sources where their users in increasing numbers are looking for solutions and looking for their products.

Noah: I it. I noticed that you're working with a ton of great agencies. saw that Animals is using the platform, which I thought is super rad in the content marketing land. Very impressive, you know, in terms of the folks that you've been able to pull on board. I looked at who's been doing webinars with you, and it's just a ton of our industry's thought leaders. It's not easy.

to get a lot of folks rowing in the same direction, much less coming on a webinar with you early on. So can you tell us a little bit about what that's like, trying to get all these people like Mike King, Kevin Indig, Steve Toth? what's it like?

Like, how are you identifying these people? How are you recruiting them? Like, must be super fun for you. I know what it's like getting people on a webinar.

It can be a lot of work or it can be exciting.

Eoin Clancy: Yeah,

yeah, it's definitely a lot of work. And when Josh joined the team, probably was that like November time, Josh, and there was probably like three or four I had under my belt and with some like amazing power users on the platform. And that started to pique different people's interest.

So and my king became familiar with us that way, as it Steve taught or like George at minutia. So people who like know their SEO.

Josh Spilker: Yeah, December.

Eoin Clancy: at least like open up the gate of like some brand awareness for us. Then we needed to ask for favors. That was basically it. like around Kevin or Eli, they had like interest in the platform and they wanted to learn more. So what we started to do was basically like co-build sessions. If someone was basically unsure of what we were trying to do or thought we were just there for AI slop or massive scaling programmatic, we were like.

let's prove you're wrong. Like let's like actually co-build something that you'll get value from. And I think that convinced a lot of people that we were thinking differently to the 1.

0 solutions that are out there. And so originally it was, we had a little bit of hurdles to overcome in terms of convincing folks. And then kind of like our sales calls, once people see the platform, see the components, basically speak to our team and see that we're trying to create.

interrogative, persuasive, cited content. They quickly turn around and they get very excited. We had a lot of amazing people on pre the new year and even since we've had some amazing quotes that we always reference internally, like when Mike King was on, again, we don't pay for any of these people, it's just we want to co-build with them and want to get their ideas out into the open.

But what was my king's line Josh? He was like, I love this product. Yeah.

Josh Spilker: Yeah,

Noah: He's kind of picky.

Josh Spilker: something really, it was like, they didn't pay me a dime, but I love this product. Maybe just to tag on what Eoin said there is, I think there really does become an aha moment when people are like, you can do internal linking, you can do cannibalization workflows, you can do all of these complex, so to speak, like tasks that a lot of SEOs and content people are doing. And it's not just about like the generation output, it really is like,

Eoin Clancy: Yeah.

Josh Spilker: using AI workflows to really stitch together a lot of the manual processes that several of us are doing.

Noah: I love it. Okay, so we're gonna jump into the platform. don't one of you guys try and get it open and we'll do some screen sharing and I'm gonna pepper you with questions as we go. and you just said something that caught my ear, of course, AI, slop and programmatic. So give us some differentiation here. Like what are those things? How are you not those things?

Get our head in the right space.

Eoin Clancy: Yeah, so let me maybe share my screen. can do a couple of things. Screen share.

All right, maybe give me a thumbs up if you can see this winning content for a yo. Okay, super. So when we think about trying not to be AI slop or promote like spam content, I won't fully go through this, but like Alex Halliday, our CEO, like personally wrote this piece and it's something that he believes a lot in and it's what he preaches internally. And we then try and like preach to customers, but

Winning content is not just gonna come from like a single prompt that you put into GPT. Like sure, if you're a small startup and you need to get your first blog live, it can help you. But when you're talking like massive enterprises or small teams that wanna get their perspective out there, like my own personal background is working with like YC companies where they need to have like three blogs on their five page website, but it needs to sell them.

I need to sell them to customers and to VCs. So here we're talking about how to get your content ranked and cited in answer engines, but it comes down to every piece of content that you want to put on every different channel. So our four pillars, and then I'll jump more into the product, are be unique and authoritative.

So you have internal context from sales calls or product information like...

I know from what we've had on Josh's side that's worked best when we share out metrics, it's when we share out like a perspective, like that's where our customers resonate and try and be persuasive to agents and humans. So just, think you were mentioning it like earlier in our prep Noah, but be it like cosine similarity, that there is like little tactics and tools and tricks that you should be thinking of. And I think Seer Interactive do a really good job on their insights hub of like speaking to these, but.

When you think about trying to rank in the SERP or rank in like an AI overview, there are different, there's different tools and tactics you need to keep top of mind to ensure that you're optimizing your content for those platforms. Yeah, again, that kind of overlaps with like ranking and existing and emerging graphs, but what you need to do traditionally in the SERP has changed. And if you want to be also ranked in AI overviews,

You need to have much stronger brand presence. comes much further back to like your domain authority and your domain presence. And then again, trying to be cited.

AIO reviews, they're not the kindest in linking off to your website, but if folks want a more detailed perspective on it, check out our last webinar with Kevin Indig from like two weeks ago where Kevin goes through like click through rates. They have dropped off, but the conversion rate of that traffic that's sent to your site.

significantly higher because someone is solutions oriented. I know for us, we're thinking about webinar software. So we go to perplexity, give us the best solutions for that include X, Y, and Z features.

And then we click on the top two links, maybe then rather, rather than go into like a Google search results. So just when you think about like AI slop, I think these four pillars kind of like go against creating AI slop.

Noah: One of my favorite examples of that, I was in a clubhouse room with somebody and they were telling me that they were capable of producing eight articles while sitting in a staff meeting every Tuesday. And I was like, okay, great. That doesn't really sound like an example of expert in the middle. Speaking of, I think you guys care about having humans in the middle and accuracy and precision.

Do want to talk about that a little bit?

Eoin Clancy: Yeah, totally. Like I'll kind of show you the three pillars of the platform. I haven't pulled up our best instance. It's one that I'm kind of like creating some workflows for training people in. like in AirOps, and let me make sure this is like pretty big, there's like three key like sections. There is your data. There is it's under home, but creation of workflows, which is where you can.

Noah: Cool, awesome.

Eoin Clancy: codifier logic and we'll show Kevin Indig's workflow in a few minutes. And then there's a grids, which is like this. I'll show grids in a second, but it starts with your data. So everything I spoke about there, perspective, authoritative, cited, it comes back to like basing or grounding your content in your best insights internally. So for a brand case, I can bring up like Airbnb.

but this is like a V1. We can spin this up out of the box for any brand. You come in, you give us your domain and we start to describe on your behalf, like information about the brand, who their competitors are, their point of view, and then starting to get a little bit more into like writing style.

So all of these are attributes we would expect you to add on to. So some companies, I presume like Airbnb will have pages of context for each of these, but this is an out of the box version.

these attributes, they can get used in your workflow. So again, like that guy who can maybe generate eight posts in a 30 minute meeting, maybe they're on a two out of 10 quality. If he was to start to integrate these as attributes, maybe he's going to get to like a three or four.

It might sound more like Airbnb, but it sure as hell isn't going to have much of their like internal data. So that's really where the knowledge base comes into play.

which in this example, I'm connecting like our site map, which we can refresh daily. And you can also drop in, what was I doing earlier? LinkedIn post templates, bring in all of your data, be it like a CSV or your live website or PDFs, whatever it is, like sales context, drop it all in here.

And then it immediately becomes referenceable in what I'll get into in a second. But I guess I'll pause there in case there's questions.

Noah: So basically in LLM land, we're grounding the data, right? We're grounding the tool to your website and your brand by putting all this data in.

Eoin Clancy: Correct. Yeah.

Let me see if I have a good example then of a workflow.

case study, lead magnet, maybe webinar to blog post. And this is one that we need to like actively update, but we were speaking about webinars. We quickly today turn what we see here of like a transcript of our call, just like we're on here, into a blog post that gets published on our website.

So basically all of the ones that you see in this section here, they start off as a transcript from a call.

what we want to use AirOps for is can we get to a good like 80 % version of a blog draft that we can then start to add the final 20 % to. So it's been a little while since I saw this one, but this is our grid. So it's the third pillar, but when you have codified your logic and you know what you want to do, and then you want to run stuff on a recurring basis like we do here once every two weeks.

And the grid is going to let you like input the like core attributes and then map all of these to your workflow. So if I do map attributes here, like it's pretty simple, but if we happen to change company name overnight, I can change the reference. I'm dynamically giving it the transcript and then just for more context and like grounding it in what we care about.

Luma is where we host our invites. If you want to sign up.

we're gonna point you towards Luma. Josh does a really good job of outlining the key points and like who should attend. We take all of that context in when we go to write the end blog post, just so we're starting from a foundation of again, our tone of voice and some context that we initially started with.

So then going under the hood, and it's been a little while since I saw this one.

Noah: This gets

us to the third pillar, right? Like we have your data, grids, and workflows. OK.

Eoin Clancy: Yeah. Yeah. And

this is like a quite simple workflow. Josh is going to show our like super powered version, but if you wanted to get something up and running pretty quick, you'd probably start with something that is on the simpler side like this. end to Yeah.

Absolutely. Is that good?

Noah: Can you zoom in to the first one? Okay. Thank you.

Eoin Clancy: I can go one more. So what we generate here, I'll actually, I'll go back one step just so you see an output, but I'm going back to the grid. So we give all of our input into this column here, which is our workflow. We run a single version and out the far side, we generate, it's coming up very large. We generate like the sketch of our blog posts. So we go through.

We want to show like a TLDR, we want to show like some of the challenges they face, the results achieved. And then on most of our webinars, we go through a workflow that we would have co-built with someone. So it's all like grounded in the structure and the format that we like on our side.

And then it starts to get enriched with the content specific to the call we just had.

So what we do is, and again, Noah, you mentioned like grounding in your data. The very first step after we take in all of our inputs is we do like a knowledge base search. Here, I just pull in one example, which was one that we like really manually wrote, but we wanna make sure that structure wise, tone of voice wise, personality wise, we like remain true to ourselves.

We're pulling an example out of the knowledge base that we showed earlier.

basically feed it a lot of our context. So at the top here, we're saying it's essentially Josh, it's a content marketer for AirOps. We tell it again, pulling from our brand case, which you can start to see if I look under the hood about your company, here's like attributes about our company.

like about, we can bring in like personas. We start to feed in all of these as attributes.

the guy you referenced who can generate eight blog posts in a 30 minute meeting, he is just going to GPT and he's just doing this lower portion here, which is saying like, go write me the blog post and maybe here is like the call transcript. What we find better, which we like advise customers to do is this like system prompt. This is just education on prompting. System prompt is you telling the LLM like,

the personality it should take on. So for all the rest of your conversation, who do you kind of want it to be? So it could be an expert salesperson, it could be an engineer.

In this case, we want it to be a content marketer who's skilled at writing post webinar blog posts. And then it's really important then that you start to, there's the notion of like fine tuning an OOM, which gets very complex, very quick. But then there's the much simpler notion, which is easier to understand.

of like a user-assistant pair. So the reason I was scrolling was user is here and assistant is here. And what you're doing in this case is you're almost mimicking or self-generating the first bit of dialogue between you and the LLM.

So what we're doing here is rather than just saying go and generate the brand new one for us, we're almost saying like, here's an old one that we ran.

And then we're telling it exactly the output that we generated and we were really happy with. And then we start to ask it, okay, great. Now go do the new one for us.

So in addition to like grounding it in the prior example, we're training the like LLM in this very simple example on exactly the input it can expect in this user part and then the output quality and type that we expect in the assistant part.

And then finally, we get it to generate the new one for us.

Noah: Mm.

I do a lot of prompting and what I'm looking at is a lot of best practice stuff. I'm seeing that you're using XML tags. I'm seeing that you're giving it an example of a good output and you're giving it the tone with which you want to accomplish that. I mean, these are all the things that one reads about how to get great outputs, whether you're working with cursor or Windsor for any IDE or whether you're working with, you know, clod or

Eoin Clancy: totally.

Noah: chat GPTF in the cloud.

Eoin Clancy: Yeah.

And just one other bonus thing I'll suggest, there's different models you should use for different use cases. the speed at which the vendors like produce models is so, quick. But when it comes to writing tasks, we generally suggest people use Cloud 3 Opus.

so that would just be like a little best practice. If you want to do stuff that is more...

Noah: Ooh.

Mm-hmm.

Eoin Clancy: And again, we'll get into it hopefully on Josh's side, like Kevin's workflow where it's a lot of data wrangling. It's like interpreting like SEMrush data or data from the SERP. Like you're just better off using a much cheaper model like 4.0 mini or 4.0 from OpenAI. And then when we think about like strategy and being able to maybe update some content live, or if you want to write like a spur of the moment, like social post in the morning.

Noah: Mm-hmm.

Eoin Clancy: like perplexity is going to give you access to the web. So just in addition to having best practices that are true across all of the models, you should choose a model that's best catered for the task.

Noah: I've actually never used Claude Opus, ever. I always use Sonnet. Yeah.

Eoin Clancy: interesting. The,

even more recently, we do find like 4.5 pretty good for like longer form content because it will like stew on the context that you've given it. It will take longer for it to give you the response, but it's like generally good when you've like a lot of information for it to digest.

It's like, it's a reasoning model. It's able to like think about that for a long time.

Noah: So and we're just looking at one step, right? This is just one step.

Eoin Clancy: one.

Josh Spilker: Yeah.

Eoin Clancy: Yeah, this overall is a very simple workflow. Like we're actually meeting in person in New York next week to make this like better so that like the, we get this to like an 80 % level, we think we can stretch it ourselves to like 95%. We want to include like videos. We want to include like other quotes and snippets. It does a really good job, but we can definitely stretch it more. And then

Noah: Yeah. Sure.

Yeah.

Eoin Clancy: when I mentioned videos, that was like, or that is a thing that like you need to include in your content to make sure that the AI overviews or YouTube or LinkedIn or Google will reference it at a higher rate. So again, very simple here, but all we do is we give what we want out of it is timestamps for where we should clip in our video. We could do the clipping ourselves. That's like what we're thinking about.

But in this case, again, we just give it a load of context and then we ask it to find, or we ask it to give us back like five to seven different key moments that share insights, outcomes, best practices. And then from that, we're able to like trim that down to like three that we care about. Go to Descript, edit it, get the file, add it to the blog post, bit of editing, we go live.

Noah: Super rad. Are we looking at a concept called content engineering? Like, when do we get into content engineering? Like, tell me more.

Josh Spilker: I would say so, yeah.

Eoin Clancy: Yes.

Do you want to give your opinion on that, Josh?

Josh Spilker: Yeah, sure. what you're seeing here is kind of beforehand, you might hire a writer or a specialist or even like a marketing coordinator to help you do some of these production tasks. And now you really need someone who can like stitch together these different workflows, think through these different models, process map out.

this workflow that you saw. And so, yeah, we're calling that a content engineer because they would also need some writing chops, some SEO knowledge, some editing ability to be able to really take this from soup to nuts. So, yeah, that's definitely what a content engineer does.

And then you can customize it to your individual domain and your workflow so that you get the outputs that you want.

Noah: When I think about you and your past roles, Josh, you've done a bunch of work with SaaS companies and you've produced a lot of content. How has your workflow changed after joining Aerobs?

Josh Spilker: Yeah.

Yeah, I'm happy to show off some of that. My workflow has changed tremendously because I'm not stitching together so many other tools anymore. I'm still using like, still going to like some rush AhRafs to like do a lot of keyword research.

We can bring in some of that, but as far as like the briefs, the internal linking, the first draft, the outline, all of that can be done in AirOps. And so it's really helped me bring a lot of that under one roof and in one place as far as the production side goes for sure.

Then you can hand it off to a writer. You can edit it yourself. You can do a lot of different options from there.

Noah: You talked about, I've heard you both say more than once, this gets us to 80%. Like what's the goal for each of the tasks? you, how does that human in the loop? Like what's the goal for human and AI working together?

Josh Spilker: I mean, honestly, or go ahead.

Eoin Clancy: I

was gonna say the ultimate goal is 100. I don't think anyone would ever say no to an agent or a workflow doing all of their work. But right now, 80%, 90 % is the goal standard.

Where most teams are, think Josh, your most recent study showed 17 % of teams.

have what they would call like solid or comprehensive AI systems in place in their content operations. I think that number is a little bit inflated because different people will say how comprehensive something is. But I think in general, the goal is always going to be as high as possible.

But where teams are really finding success and it's going to be an upcoming topic that my king has at New York SEO week and also at like Saster.

but leading teams are basically finding the best success in like hybrid workflows. So when you see stuff on maybe LinkedIn that's like fully agentic and it's like you write in one line, you're the guy in the meeting who can generate a blog post, but he can also fully optimize the site in 20 minutes. You have to be like dubious of that.

So, yeah.

Noah: Take us through that concept of agentic workflows versus hybrid. Because when I saw your tool the first time, I didn't know what an agentic workflow was. And I was like, I finally see it. And then you're like, no, that's not quite it.

Eoin Clancy: you

Yeah. So agentic, I think can mean a lot of different things to a lot of different people. But in my head, it's, I describe the task I want done. The system then can start to recruit all of the workflows and the different tools it needs.

And it can figure out how to stitch them together and then like give me the output. So there's like a lot of reasoning, a lot of external API calls. There's a lot of structuring going on there.

I think.

across industries were well away from something that is like at that level. But what is more than achievable, and you've seen a hint of it there, and Kevin Indigs is again a greater example, but like the hybrid situation. one thing I didn't show there was like human in the loop.

It's like, it's part of our product. That's how important it is. But when we think about like hybrid, it's your thinking of the task, your like...

designing the workflow, you're going to build the workflow. And as part of that, just like on any team, you want to have like an editor process or you want to have like the brief review process. Like very much how we think about workflows on our side is trying to codify your existing processes and then really trying to use human input or human insight or expert review at the moments where it's exactly required. So there's going to be some tasks that

humans are gonna be consistently better at, which is the final mile. So, like, I wouldn't want anyone else but Josh being the final reviewer on that blog post to make sure videos are really good, they're on point, it like expresses the key points that like an SEO person or a content person would wanna know and then publish it. But I would trust, sorry Josh, would trust perplexity over Josh to be able to do like...

six months worth of research in like two minutes on like four or five different topics in prep for that webinar. So Mike King is big on like his rag system. Might not be something everyone is familiar with, but we have like another workflow that can help us do research in advance of a webinar that gives us the key points we need to know.

And we can start to orient that towards our audience. So there is a lot of us doing stuff in addition to workflows there, which to us is hybrid.

generally for the best teams are having success, we will over time move more towards agentic. The lines are very blurry, but I think as we get more, as these systems become more part of our day-to-day process, it kind of becomes agentic as well overnight.

Noah: It's interesting because when I think about what that possible future looks like, I think about like all of the ways in which coding with AI goes wrong and how you actually have to know how to code in order to like, like vibe coding only gets you so far, right? So it's like, if you have this system building itself in an agentic fashion and it builds it wrong, you're not like an AI expert and you have to like debug it and get it to work.

and you're just like a content marketer. like that's where we're getting in engineering, right? Anyway, Josh, do you want to take over? I think you had some stuff you wanted to share.

Josh Spilker: Right.

Yeah.

Noah: I didn't mean to cut you guys off by the way, I'm loving this.

Josh Spilker: Yeah, so everyone

see the AirOps cannibalization preview grid? Is that right? Okay. So this is what the workflow is that Kevin Indig shared on his growth memo and on LinkedIn.

And so I ran it a couple of times, but first I'm going to show you the output. So I don't know this company, Third Space Learning. I ran this before we started, but this was actually submitted to us.

So I'm going to show you the output of the report.

Eoin Clancy: Yes.

Josh Spilker: So they entered in this URL with the keyword of math's tutoring. believe this is maybe in the UK. And so they had some key findings here for potentially competing pages. None of them qualify as high risk of being cannibalizing each other. And so then we have this heat map grid of risk levels between this page, a guide to implementing math tutoring at your school.

This could be a medium, free KS2 maths. All of these are just pages from their site. And so then we have our definitions here of high, medium, low similarity.

And then it went through and ranked each page based off of that high, medium, low assessment matrix. So this page, medium cannibalization. You might want to take a look at it.

The target page has 6,000 words. This competing page has maybe close to 27. So 2,700 is something to look at there.

Is there a target audience overlap? Probably. It seems like they're serving similar audience. Here's another one at surfaced, low cannibalization risk.

You can also see some of those key stats. And so I found four different ones for this. So I thought that was a really good example there. And then let me give one more.

This was from, I wasn't quite familiar with this. It seems like an Airbnb in, was that Denmark or Germany, I guess, with the DE.

Well, let me bring up this example. So they entered in this to see if any pages were competing on the keyword for this right here. And so they said, or the workflow AirOps came out with, there could be five potentially competing pages, no direct overlap.

And then once again, you can see like low, low, and then another low.

So this was just an example of like kind of the final output that we got that like was built from scratch with Kevin and Nicole and our team. So now let me get to what everyone really wants to see is all of these different steps. There was more than I believe 400 steps as we were kind of going through it.

Yeah, 380, 382. They're starting to build up. This is a power step, I believe. Is that what this is Eoin? Yeah, these are brand new.

feature that we just released this week where it kind of shortens out some of that. And if you want to see like all of the information in this, for example, we have this toggle button down here that will then show you, for example, here was the prompt in chatgbt 4.0.

You can see all the Python code all the way down.

Eoin Clancy: we just see

your grid, Josh. You even need to reshare, sorry.

Josh Spilker: you just see the grid?

There we go.

Eoin Clancy: There we go.

Yeah.

Josh Spilker: So here's what I would say with the, can see it step by step, but then you can expand it out and you can see the actual like Python and Chatchity there as well. And a really great feature is our new copilot. So one thing you can do, this helps you build along the way and gives you advice and tips and helps you troubleshoot. But a nice button we have here is just explain the workflow. So if I click that, it'll give it a second because this is a very long workflow, but it'll show you.

all of the different steps and how it comes together.

Eoin Clancy: I think the other thing to add here, do want to just zoom out a little bit, Josh? Just so like the density of it. There's like a lot of different steps. like this is Kevin and digs like idea, which we keep referencing back to. It was built by Nicole on our sales team and Nicole has, from my best knowledge, no exact programming experience. And like Josh, they're referenced multiple. It's right there in middle of the screen. Multiple Python steps.

Nicole has no coding knowledge. like everything you see here is her using the co-pilot on the left. And then also in the top right of Josh's screen, you'll see there's a little generate with AI.

So be it for a prompt. If you do not know how to best set up a prompt, you just ask that little like button there for code steps as well. If you were to drag one on, you can just say, I get data that looks like this.

X coming out of this step and it'll do a lot of the work for you. then, just like Noah was saying with like vibe coding, you're not gonna know exactly what it does, so you might get an error. So what you can do then is the copilot works with you, where you just say, copilot step one isn't working as I expect, it'll go through all of the logic, all of the errors, and it'll like give it back to you. So you can come in with like a very good strategy and just know that you need to like,

go end to end with it and just the system will work with you for complexities involved to help you figure out the nuances.

Josh Spilker: Yeah, exactly. And then on the co-pilot side, you can hover over each of these, extract domain, go to the prompt, and then you can see the content analysis section. Find the keyword. So if I just click on that, it'll take me straight to the prompt SERP overlap, for example, and then it will dive in straight into that as well.

And then it competitor and then the report generation. Here's the sentiment level down here showing you exactly what that code steps are for this. And then the output.

And then we do publish to the grid, which is that section I showed you earlier. So you can run like this off of multiple URLs all the way. Let me just show this.

you could, mean, all of these are different companies because we had different people submit on this grid, but you could have...

10, 20, 50, 100, all of your URLs listed here, run this exact same analysis if you wanted to, all the way down, and then have an output like this to see where they would overlap or compete against one another.

Noah: I see this as like the perfect opportunity to have a professional services team. Like this is a perfect way to make your tool sticky and to have your professional services team provide value. mean, like some of this stuff, like getting vibe coding solutions would be totally fine, but other things it's like, ooh, that just made this step to an infinite loop that...

Josh Spilker: Yeah.

Noah: Like an infinite API requesting loop or hitting some rushes API. You know.

Eoin Clancy: Yeah, to

add to that, like, our new terminology, like internally is like the likes of this is a skiing reference, like a double black diamond. You are not, if you are not familiar with the AirOps today, unfortunately, you're not going to come in and achieve this yourself overnight. But when you think about like the bunny hills or like a green or a blue sort of like track, you can create if you wanted exactly this.

aversion that probably again, like my phrase of the day is like gets you 80 % of the way there and like pretty quickly within like one, two hours. So again, like what we sell to our like larger customers are those professional services. It's like you get the Lamborghini, you get the Brigade Veyron, but if you really just want to go like down the street and you can like do a lot with that, you're probably going to be fine with your Nissan Micra or Ford Focus. So

Like the level of complexity again kind of depends on what you're looking for, but all of these solutions, you can start to strip back the content and the complexity to achieve what you want. And what I will say with this one, it was not designed with that many steps in mind, but the framework never changed. initially I think Nicole's workflow was like eight steps and it was like very understandable.

We might not have needed to ask Copilot to explain it to us. So.

There are a lot, can go a long way by doing like stuff that is like pretty simple, like in the platform. When you want the Lamborghini, we have the content engineers internally to help you out.

Josh Spilker: Yeah, I can show actually a different one as well.

Noah: Yeah, we have time.

Josh Spilker: This,

for example, is a content brief. Let me show the output first, actually. So this is actually live from something I've been working on. You can see my different article topics here.

So if I want my brief, I'll do organic search marketing, for example. That's what the raw looks like. But then I have it in the column right here, where it has the role of that.

Noah: We're looking at the flow, the content brief, JS flow it looks like.

Eoin Clancy: Yeah.

Josh Spilker: So, oops, sorry,

Eoin Clancy: There we go.

Josh Spilker: missed it again. Here's my content brief off of the grid for organic search marketing. example, this is just some words I'm playing around with, some keywords. You can see like my title, my title, topics to cover, search intent.

Noah: No worries.

Josh Spilker: related questions from Google. And then it has suggested internal links. And so this is a content brief that I can start with, basically out of the box in AirOps. And then I wanted to show. So that's what the output looks like. And then if I share this tab, you'll be able to see all the different steps, not near as complex as what that cannibalization workflow is. If I click here.

you can then start to see like it dives into the SERP API searches, the knowledge-based search, which is the AirOps knowledge base. What you're looking for with search intent, it prompts the LLM. I did a web scrape step here.

You can also do a knowledge-based step, searching what's been indexed for internal leaks gives me title options off of from chatgb4. then, yeah, even.

gives me some potential anchor text if I want to use that all the way down the line. So it gives me a brief that I can then hand off to a writer, write myself, whatever I want to do with that, or have a first draft from AI, another workflow as well.

Eoin Clancy: And I think one of the things just to call out Josh, if you want to stay on that screen, like if you go up to the internal links portion, like if you wanted to get started and add significant value, like overnight, like two of the simplest things you could start with one is like bring in all of your existing blog posts and you can just connect your site to our knowledge base to do that. And you can just see going across like a very simple workflow, all your blog posts.

Josh Spilker: Yeah.

Eoin Clancy: What are your opportunities to increase internal linking? That's going to start to give you much more juice in the interim. And then even some of the biggest companies we work with, a Lightspeed or a Descript, they all start with improving their titles and their meta descriptions. It sounds basic, but it gives them a huge boost. So again, Josh had that above, but that's like a one or two step workflow that can...

It's a bunny hill, but it'll get you significant value in probably 20 minutes.

Josh Spilker: Yeah, and so for instance, all of this can be customized. You know, right now it's pulling my brand kit. You can go more in depth on here, only pull URLs that are in a certain topic or in a certain folder for internal links. Like you can customize all of this down, which a lot of these other tools out of the box, won't be able to do that.

Noah: Super red. I love this.

We talk about working in a startup a little bit. We talked about it at the beginning. We talked about what it's like when things go right and planning for success, right?

When you have, hey, we grew so fast that we broke shit. Those types of problems. But what's it like when you start to feel like you're getting traction in the market?

Can you talk about that a little bit?

Eoin Clancy: Yeah, A, it feels good. It's the best, like, sort of feedback we can get. And I think this week has really shown us, like, how far we've come. Like, even before I jumped on the call, we have, like, an internal WINS channel where we start to share out, like, quotes from people who, like, unsolicited, just say they love us or they love, like, a feature or they had, an amazing outcome or...

what we're increasingly seeing, and it goes back to your question, Noah, on like content engineering, we're increasingly seeing people who have tinkered with AirOps in the past. Now because of they can show one or two workflows to like a new employer, they're getting like a step function of like a promotion either in the same company or at a different company. So in addition to like giving outcomes to amazing people and teams, it's really cool to facilitate this new, or guess like revamped job title.

And it seems to be like we're teams and individuals are looking to invest in that given skillset. So at a startup, we move and adjust very quickly. Even the term content engineering, think Josh was like, it might've been something else up to like an hour before we published the original blog post.

So some things are, you have a good feel and you go with it. But I think the best form of feedback we get is directly from users. So we did a big push Monday for 2.

0.

And just a lot of the feedback we've gotten since just via DMs or on Slack or on LinkedIn, it's helping people move quicker and better and faster internally, and they're getting great wins. So I think that's the best feedback we can look for.

Noah: As you were talking just now, I started to think about the value of having an Academy and having like an AROP certification and having like, you basically control the narrative of how people will be successful with your tool and having that success be something that people would hire for. And it just sort of like builds the loop a little bit.

Josh Spilker: Yeah, mean like HubSpot did, go ahead.

Eoin Clancy: We'll give a little teaser. yeah.

Yeah. I was gonna say just, like Hub Swap did, we'll give a little teaser here today. So we do have upcoming cohorts where we'll be training people on the platform.

So that'll be like live cohorts. You work with our team to go from, you have some knowledge of prompting or from knowledge on the platform or a different tool where we'll over the course of a few weeks bring you to hopefully that.

double black diamond hill in terms of complexity. But in addition to that, we're investing a lot. We just got our first version live Sunday night. So that is a startup.

We got our first version, our new version live of AirOps Academy. So if you want to get a teaser for a lot of the pieces, myself and Josh went through there in the grid and the workflow builder. We have a collection of like 10 or 15 videos up there right now.

They'll give you a starting point again.

where people want to make it their full profession or really want to upgrade what they do internally, we will be running cohorts. Hopefully very soon you'll hear more from us.

Noah: Can you let us know about some of the different APIs that you can connect to out of the box?

Eoin Clancy: Yeah, I can pull it up.

Noah: set of workflows,

thought that was one thing as a builder I always am excited about. I know you can connect to Search Console in SEMrush.

Eoin Clancy: Yes. Maybe Josh, you run through that and I'll get a little screen share off and running.

Josh Spilker: Let me find our list.

Eoin Clancy: So here we go.

screen, entire screen, here we go. So if we wanna drag in a new step, when we talk about vendors or different APIs, you can do everything from integrate directly with Google or do search steps with our own scraper or perplexity. We like more out of the box custom stuff like code steps, but really it comes down to image generation, so.

We work directly with like Getty and a few other, like template image providers. And then we have B2B enrichment. So when you start getting into, you want to personalize your like inbound marketing emails or like your, your signup flow.

We see people do a lot here with like brand fetch to get like a logo or get some like overview of company. You can use Hunter, people data labs to get.

demographic information for SEO research. We have data for SEO, SEMrush, and then a couple other vendors coming pretty soon. So these are all like out of the box.

Like if I drag in one of these steps here, what happens, sorry, if I drag one of these steps here, what happens is for this one, I want to get the domain organic search pages. I would just give it a domain. I would tell it how many results I want back.

And it's just going to give me SEMrush.

info. So what is really nice here that I think we can do a better job of telling people about is you don't need to come with an API key. You don't need to have SEMrush internally.

So that means that rather than needing to buy multiple tools, if you've just got AirOps, it comes out of the box with a SEMrush.

I think we have it down further, but with like Getty images or some of our other providers, you're going to pay like pennies on the dollar versus what you would pay them if you were to go natively to them. So content quality, like evals, have originality.ai and then we have a lot of companies who obviously they produce content or they have content stored in other systems and they want to bring it into AirOps or maybe they want to do their content review system or

some other part of their process externally. So you can always ship it to like Google Docs, Sheets, Notion, some massive e-comm customers run their entire store and their operations behind it on top of AirOps and then ship it all back to Shopify. So there's lots of integrations there.

Noah: You've got a whole bunch of new CMSs. I remember the last time we looked at this, you just had Webflow. I mean, that's my memory. In just a couple months, you've added the rest. That's pretty rad.

Eoin Clancy: Yeah.

Yeah.

And like what we'll say as well is I know people have like a diversity of CMS's and to integrate with your CMS, if it's not provided here, again, this copilot is your best friend. You don't even need to give it like docs per se. But if you work with, I don't know what another one, but like some Adobe solution, you could say, I want to be able to like send my content that I've generated here to this.

like workspace and it will help you generate again the code to do that for you. So it's a one time thing. You will generate it, you'll lock it away and then it'll just live kind of abstracted here as in like a right to Adobe step.

So that people want help with always reach out to us obviously, but that is possible again where we don't have it like out of the box.

Noah: Tell me about support. What's the support like Monday through Friday? it seven days a week? Is it 24-7? Is it 8-2? 12-1?

Eoin Clancy: To your earlier question, we're a startup. we're 24, 7, 365 or 366, including leap years. So we have like intercom. So when you're in the platform, the little chat bubble down the bottom, right. If you want live help from a person, you can make requests there. And we get into like our larger like contracted customers. We work directly with them over like Slack channels. So anytime a day they have like a direct point of contact and usually like

Noah: Yeah. Yep.

Eoin Clancy: our engineering team or our product team is jumping in to help them as well. for anyone who's on a call today or interested in getting going, there is always still someone behind the little intercom bubble. And I think our time to respond is like under 10 minutes. So basically, there'll always be someone there to help you.

Noah: I found when we were building a product that pricing was really hard. Have you guys found that to be the case too? Like getting it right for different size, know, right sizing it for different fits is tough.

Eoin Clancy: Yeah, it is tough. What we've learned though is there's probably more value that we offer customers that we've like yet to capture. So I think there's like, there's definitely like some right sizing that we're going to do over the next while. But in terms of like delivering value to customers and customers getting an amazing experience, that's generally what we're trying to prioritize right now. So someone comes with like an opportunity or

some problem they've had and like drop off of traffic. We're really trying to like right size so that they get the best outcome. And then again, when you work with our team, just like what you mentioned on like the professional services, like every one of our like contracts will come with someone who's like dedicated to get your solution stood up.

yeah, customers are seeing a lot of like immediate value. People should go and check out our most recent customer story with Descript.

but Sung, is their acquisition or like growth marketing acquisition manager, he said the time to value with AirOps was like, blew them away. And that convinced them to like want to proceed and like up level the more use cases they did. So we have a sales experience right now that's like optimized to delight customers.

We'll worry about pricing in the future, but we're thinking about it.

Josh Spilker: I'd say dscript was a great example because I know they're really respected. At least I respected them from just their SEO presence and their organic presence. And our team worked really closely with them on just their content refresh from meta titles, meta descriptions, doing some kind pre-work before this workflow on cannibalization and duplicative content. And our team really knocked it out of the park for them. It's a really great example of what you can do with AirOps, for sure.

Eoin Clancy: Like when we think about the sales process, it's our only thing about the post sales process. It's almost like a sales process again, in terms of like detail and discovery. So just like what we went through there in like the brand kit or the knowledge base, like when we get into someone's workflow, like the nuts and bolts, we have to think about like how to ground it in their data, making that system like really reliable. We have to think about like integrating their tone of voice. So there's a lot of.

context that we like work with a customer on trying to collect and then that's what really like stands up the solution to be our double black diamond or like an amazing outcome versus something that is closer to the guy who gets his eight blog posts in the meeting.

Noah: Wow. Well, this is an amazing time for us to kind of wrap it up. I just want to say thanks. You guys are amazing community partners for the SEO community. you're the first community partner that I've had on a campfire chat. And really, it was fueled by seeing the steps and thinking to myself, my god, that's a data pipeline.

You know, like, my God, I got this, I can do this, you know? And I wanted people to be exposed to, at the time I thought it was an agentic workflow, but it's not, it's really a hybrid thing. And I think people's mindsets need to shift towards like what next looks like.

And I think this is definitely a compelling version of it. I'm just super grateful that you guys came.

I'm also grateful for everybody who's watching. Some folks have mentioned that the live stream has been super jittery for them, which is wild because it's been perfect for us in the studio. But I'll edit it and put it up on YouTube as quickly as possible so they'll get a great cut of it.

And just super, super grateful that you all joined us. Anything you want to leave our audience with? Anything you want them to do after the chat?

Eoin Clancy: If folks want help getting started, find myself or Josh or one of our teammates on LinkedIn. Like on our website, we have like a book a demo slash. If you mentioned like where you found us as like the SEO community, one of us will personally look after you and make sure that you're taken care of in terms of we can drop some extra templates in your space, give you like extra enterprise access. So if you do come through, do mention how you heard about us and mention the SEO community and then just give in.

Coming on Josh's end, bit of promo for next week. We're doing a session with Seer Interactive. Will and Alyssa on their team. If you want to learn a bit more about what we're talking about is not exactly Aerox.

It's how to like think about these terms like AI overviews, AEOs. If you're an SEO, how can you think about communicating those internally? So that's the core of our focus, which should hopefully help a lot of you as you're trying to like advocate for projects and budget.

like kind of manage up as well. So that's the focus of our conversation and we'll want to listen to a lot to say. So we're excited for that.

Noah: anything from you, Josh? Any parting thoughts?

Josh Spilker: We do have a builder community coming up soon, like you mentioned, to equip people to be content engineers. And also at the end of April, right around SEO week, putting out a report called the State of Content Teams, just some survey data that we did about how teams are thinking about AI, how they're structuring their teams, and how they're adopting to some of these new workflows that we're talking about. So that would be a really good research report as well.

Noah: Awesome. Well, I'm going to stop the recording, but you guys stay with me in the studio because we got to finish doing our uploading and stuff. for everybody at home, thanks so much for joining us. This was another amazing Campfire chat. And talk with you in two weeks when we get to hang out with Jamie Indigo, which will be super exciting. All right. Take it easy, everybody.

Eoin Clancy: See ya.