Information gain is the new knowledge a piece of content gives an AI model, and Christian Rigg, Director of Strategy and Operations at Eleven Content, argues that the best source of it is already inside your company: meetings, team chat and internal docs. He tells Paris Childress how Eleven built a governed knowledge base that traces every insight to its source while human writers still do the drafting. The conversation also covers how much content is enough, where SEO and GEO overlap, why digital PR shapes the AI opinion of a brand, and why sentiment matters as much as mentions.
Most companies already produce the raw material for original, expert content every day, in team meetings, chat channels and internal docs that marketing never sees. Christian Rigg, Director of Strategy and Operations at Eleven Content, a London-based content and marketing agency, explains how Eleven captures that expertise in a governed knowledge base, why its writers are still human, and why GEO now depends on what an LLM thinks of your brand.
Because AI answer engines, like Google before them, reward content that is original. Christian borrows Paris's metaphor of LLMs as truffle hunters: to be surfaced, content has to stand out from the crowd. Paris adds that experience and expertise have to come through in the content itself, while authority and trust also depend on other sites mentioning your brand. He describes information gain as the degree to which a page gives a model information it has not already trained on.
From inside the company. The problem is that subject matter experts are time-poor. Eleven separates two kinds of expertise:
A couple of years ago, Eleven concluded that the best information gain comes from within the company, so it looked at where people already talk: team meetings, Slack and ClickUp channels, and its own Industry News channel in ClickUp, where colleagues drop a LinkedIn post and argue about it.
"Not the writing, not the thinking, but getting the thinking into the writing." Christian Rigg
You collect it in one organized database and let an LLM work across it. Eleven's tool takes meeting transcripts, Slack and ClickUp channel information, documentation, voice notes, interviews and existing marketing content, and stores it in BigQuery on Google Cloud. An LLM such as Gemini or Claude then breaks each company's material into themes and narratives. BigQuery lets Eleven tag every item with the date, who was involved and the theme, which is why Christian prefers it to Google's File Search Tool, a managed RAG system.
Paris's agency does the same with a Google Chat project space for every project, with direct messages always excluded.
By being strict about what goes in. Christian gave governance a whole slide in his brightonSEO talk:
Manual curation limits the volume, Christian says, but makes for a better database. And tools like Slack AI and ClickUp's AI already run on company data in the cloud, so a company might as well build a system it controls.
At Eleven, no. Human writers draft, because the best content comes from people who understand the subject and can write with nuance, and Christian says Eleven has seen higher engagement and conversion when readers feel they are hearing from another person. The knowledge base is used for briefs. Paris takes a different route at GEOforge: the first draft is AI-generated, but only from a brand's own knowledge base, and a human then fact-checks every claim. Both agree that the source of originality matters more than who types.
"Who cares who's typing, right?" Christian Rigg
There is no fixed number. Christian's line is the most content you can produce before it becomes derivative and repetitive, which is high for a data-rich enterprise and lower for a smaller company. He favors a few key assets, such as e-books and white papers, with supporting blog articles, LinkedIn posts and newsletters, and borrows the idea of being relevant and relentless from HubSpot's Loop.
"Don't do 50 blog articles. Do three e-books, 10 blog articles, LinkedIn posts, get on Reddit." Christian Rigg
Because LLMs express opinions. Outside technical areas such as crawlability, Christian sees a big overlap between SEO and GEO. The difference is degree: in SEO, PR was mostly about links, while in GEO it shapes the narrative when an LLM mentions you, which is why Eleven has added a digital PR arm. Paris adds that authority now flows through co-occurrence and corroboration across many sources.
Sentiment matters too. Christian has noticed ChatGPT and Claude flagging recommendations that come from a vendor's own site, and LLMs will sometimes say a product probably isn't for you. Being mentioned with a caveat still beats being absent, but sentiment matters as much as mentions. Because the buying journey has collapsed into chat, visitors now arrive with an opinion that may be borrowed from an LLM. Christian cited conversion rates of about 3% for Google visitors and 13% for ChatGPT referrals, in line with Marie Haynes's GA4 analysis of 100 sites (13.8% vs 3.02%), which also found that Google organic still drove far more total conversions.
"So what's the LLM's opinion of you?" Christian Rigg
The conversation closes on a warning. As personal AI agents earn more trust, people will hand them more decisions, and both agree that critical thinking is worth protecting.
Paris: Hi, everybody. Welcome back to The GEO Show, brought to you by GEOforge, full self-driving for AI visibility. Today, I'm thrilled to be with Christian Rigg, who is a director of operations and strategy at Eleven Writing, an agency based in the south of France. Christian, welcome to the show.
Christian: Thank you very much. Thank you very much. Small note, we actually recently rebranded to Eleven Content, because we've, significantly expanded our service offering, and we're more of a, full market solution now. So SEO, GEO, digital PR, design, audio, visual, the whole gamut
Paris: Got it. And you just told me that like literally two minutes ago, and I still it still came out wrong. I was,
Christian: It's
Paris: Glancing at the LinkedIn profile
Christian: There's still some old assets that still say, Eleven Writing. Been, I've been working hard to try to hunt them all down and update everything, but as you can imagine, it's a big job.
Paris: Yeah, that actually sounds like potentially a GEO issue. Have you all sorted that out with AI?
Christian: Yeah, yeah. Actually, it's done a pretty good job. The last time I searched us on, Claude and on Gemini, it did say Eleven Content in parentheses, formerly Eleven Writing. So it seems to have understood the
Paris: Great. Yeah. All right. Well, now that that's out of the way, Eleven Content, I think that sounds, it sounds better and more holistic. So Christian, you come from an interesting background, not in marketing. Could you just take us through your background and then, and how you eventually landed in the role that you're in now?
Christian: Yeah, yeah, for sure. So I think like, yeah, a lot of marketers, I kind of, found myself in marketing after a time. Originally, I'd done my undergraduate degree in, psychology, and I still find, psychology and neuroscience and kind of human behavior really interesting. And then, after that I spent some time in hospitality and, hotel management. And then I pivoted to do a master's degree in archeology.
Christian: Again, kind of this, inquisitiveness with like human behavior in the past. And that actually led me to, translating because I was doing my degree, here in France where I live now. I was doing my master's degree in French and archeology, when I started kind of translating free time, freelance. And then that got me into content writing. And that's how I found my original position at what was then Eleven Writing, as a content writer.
Christian: And then over the last, six or seven years, I've, helped the company to, grow to where it is, and I've taken, more of a leadership role, kind of bringing together all of my experience in management and psychology, less so archeology.
Paris: Yeah. Well, I've always believed that psychology is really at the heart of all marketing, so I think that degree really serves you well
Christian: Yeah, yeah, fundamentally. And all of my, you know, my favorite courses and everything was on social psychology, and how, you know, how we influence one another and how we behave as a society, including as consumers. So a lot of it feels, yeah, quite transferable to what I do now.
Paris: Great. Great. So Christian, your talk at brightonSEO, I think it fell at the exact same time as mine, so we were maybe competing for the same conference audience. You were in the auditorium too, and I think you packed the house there and did really well. Your talk was about E-E-A-T in a GEO-first world. And, just for our listeners, E-E-A-T is a concept that emerged in SEO several years ago.
Paris: I think originally it was just EAT and they added another E, but it's experience, expertise, authoritativeness, and trustworthiness. And this became, widely w- this became widely adopted as a sort of a checklist for content quality. And Google also was the one that really supported this acronym and gave that guidance. And my talk was actually s- on a similar topic really, which is information gain, and I think there's a lot of overlap there. But why don't you first tell us about how you see E-E-A-T in a GEO-first world?
Christian: Yeah, for sure. So I'm gonna I'm actually gonna, borrow a metaphor that you used, in your talk because I think it's really apt, and that was this idea of LLMs being truffle hunters. I really liked that, idea, about, the requirement to, stick out from the crowd if you wanna be seen. And it's obviously something that was and remain, you know, continues to be really important for SEO, and it's, you know, small wonder that Google went so hard on E-E-A-T, I think not just because they wanted to advise people on how to do well on Search, but because they recognized that for their model to work well, for their systems and their business model to work well, they needed people to be creating, you know, interesting, unique, original content.
Christian: And it seems very much, at least from what we can tell now, that LLMs, respond to that. And I think it's very much the same as, you know, Google shaping its engine to find what it knows humans are really interested in because we like the surprising and the unique, and of course, in a business setting, we want to have an edge, and you gain an edge by having knowledge that, you know, other people don't.
Christian: So I think if we could look behind the curtains of, you know, OpenAI and Claude, I wouldn't be surprised if in the, you know, the system messages that accompany all of the queries that we send, you know, they send instructions to their LLMs to find interesting, unique, original content and to surface that over, you know, maybe the more mundane or banal.
Christian: So I think it's much a reflection of these companies and these models reacting to what humans look for. And GEO, I think, has gotten us closer as a proxy to human behavior than search. You know, search had keywords, and that kind of serves as a proxy for human interest, human behavior. And now LLMs, by virtue of the fact that they, you know, architecturally, they're constructed from human knowledge, right?
Christian: Like editorial, you know, books and white papers and e-books and everything, like everything that we put out there. I think it's very natural that uniqueness and originality and high effort and E-E-A-T are ways that you stand out in those things.
Paris: Yeah. Yeah, E-E-A-T is still such a pillar, and I think of now in the GEO-first world that the first two E's of experience and expertise is what needs to... It needs to really come through in the content. And then the authoritativeness and trustworthiness, while I think that also naturally is reflected in the content, I think that alludes to some of the offsite, off-page work of citation building and brand mentions.
Paris: So being authoritative and being trustworthy means also that other trustworthy sites are mentioning your brand. And I've, latched onto this concept of information gain, and I think information gain is very similar, although it's, the audience for information gain though is more of an AI, let's say an AI bot or an AI crawler, whereas the audience for E-E-A-T is a human reader who's gonna read that content.
Paris: But the concept are similar and information gain is, our way to try to describe the degree to which a piece of content is giving a model information that it has not already trained on essentially, so net new information. So it means that probably the stuff in that article hasn't yet been surfaced online, it hasn't been discovered or published, and it represents fresh, new proprietary knowledge from the brand, and I think it's very close to expertise, experience and expertise.
Paris: And one of the challenges that we often have with clients that they buy into the whole theory of this, but they struggle t-to really deliver enough of the experience, the expertise. And so m- yeah, maybe they have some published case studies and from time to time maybe some experts will contribute to blog posts. But how do you try to help your clients to bring out more of the experience and the expertise when there's not, there's just not a ton of data lying around that you can leverage, and basically people need to either get in front of the camera or start recording s- their expertise?
Christian: Yeah, yeah, for sure. I mean, it's definitely, you know, it's a fundamental problem that we, you know, really, struggled with. And I think that, you know, all agencies or all in-house marketers struggle with the fact that most of the thought leaders in the company or many of the thought leaders in the company, are super time poor. They have lots of demands on their time.
Christian: A lot of those demands are more, you know, immediately urgent than, you know, sitting down with marketing and talking about, "Oh, well, I saw this, you know, LinkedIn post, and this is what I was thinking about it." you know, it's gonna be, financial and operations and these kinds of things that tend to be more, immediate. And so originally So Eleven has kind of two ways of addressing this.
Christian: We call it, so, we have owned, expertise and sourced expertise. And owned expertise is all of the expertise that a company owns and generates itself. It's the individuals that it has, and the data that it has and, kind of, you know, their opinions and all of that kind of stuff. And then sourced expertise is w- partly a service that Eleven provides.
Christian: So we have a very large and growing database of subject matter experts that we can that we've made connections with and that we can draw on for insights and articles, for quotes, for, you know, citations, these kinds of things. And then the writers that we use for the actual content also have expertise in their background. We, you know, we, if it's, HR, for example, we have a couple clients in HR, then, the individual might have a master's degree in HR, and they've worked in HR for the last, like, five or 10 years, this kind of thing.
Christian: So the source expertise was always something that we could provide. But, you know, a couple years ago, we realized like so many others that the best information gain comes from within the company, you know. And that's how you demonstrate the company's authority, trustworthiness, experience and expertise. And so, we set out to find ways to make this, you know, easier for ourselves basically, because pulling that out of companies can be quite difficult.
Christian: So this was the topic of my talk at brightonSEO, and I won't redo the whole thing here. But, essentially, it's a fairly like technical, solution to this problem, that came about from us trying to think about, okay, where are all of the places that people just naturally talk about these things? Where are all of the Like, where do the interesting conversations happen in our company?
Christian: That was really our starting point. And, well, I mean, that's team meetings. It's Slack channels and ClickUp channels. We have this one really glate great ClickUp channel called Industry News, where every day we you know, I see something on LinkedIn, I'm like, "Oh, that's interesting. I wonder what, you know, my colleagues will think about this." And I drop it in there, and then we all have like a flurry of discussion.
Christian: "Oh, I don't agree with that at all. This is completely wrong." "Well, have you thought about this thing?" and so all of that is, you know, thought leadership and experience and expertise. And it's true that, you know, data is a fantastic, way to demonstrate an experience and expertise. But thought leadership can also be expert commentary. It can also be opinions, you know, op-eds.
Christian: It can also be your take on things. And very often what people want from, I'd say individuals within a company, right? That a company generates data, that's really interesting from a company perspective. And then the personalities at the company, which is very often what really sells a product, right? Is the people at the company.
Paris: Yeah, especially with B2B, I think. Yeah.
Christian: With B2B. Yeah, exactly. Especially with B2B. You know, you're looking for, somebody who is, expert, opinionated, worthy of your time and attention, you know, whether that's on LinkedIn or in trade media or whatever. And so finding ways to extract that from thought leaders, time-poor founders, et cetera, was really important to us. And so we basically developed this tool, that in a very safe and secure way, can, take meeting transcripts, take, you know, take, Slack and ClickUp channel, information, documentation, voice notes, interviews, existing marketing content, put it all into a, organized database, and then use generative AI, so use an LLM like Gemini or like Claude, to then be able to work with that.
Christian: And we identified a number of different, you know, for each company, we break it down into themes that matter for them, narratives, that might play well, for them on LinkedIn or, you know, on blogs or in eBooks or whatever. And it's honestly made it significantly, easier to take what would normally be kind of like a, you know, a blog post where we really struggle to say, "Okay, what's fundamentally unique and interesting about this?" and it's actually helped to kind of, you know, almost automate that process.
Christian: Not the writing, not the thinking, but getting the thinking into the writing, which I A lot of companies really struggle.
Paris: Yeah. I love that. Really compiling a knowledge base, and when you think about it, that knowledge base which is pulling in meeting transcripts, Slack channels, ClickUp discussions, ClickUp tasks and projects, that actually has so much value to the organization even beyond marketing when you think about it.
Christian: Absolutely. Yeah, yeah. I mean, we, you know, at Eleven we're always looking for better ways to serve our customers. We're always looking for better ways to do what we do. We wanna stay, you know, really on We try to walk a fine line between being on the bleeding edge of what we're doing, but taking a very measured and reflected approach to everything.
Christian: And so it's very useful for us as well to, when we get together as a group talking about things, then we can access this database and say like, you know, what has been our kind of universal consensus on this topic over the last six months, and how has it changed? And just having access to that is very valuable for us to be able You know, continue to serve our partners
Paris: And architecturally, how d- how what does that look like? Is that a Claude project, or is it something bigger? Where are you actually storing this data? Where is the database?
Christian: Yeah. So it's in BigQuery. Most of my kind of development experience is in, Google Cloud. I know some people, you know, prefer, Amazon Web Services. I don't know anyone that prefers Azure, but some people use Azure. But,
Paris: We're really big on Google Cloud. Actually, our, All of our database is running Google Cloud for GEOforge, and I mean, generally it's very reliable. And I think you can trust Google. In the whole world of AI, I know a lot of companies are still not that trusting of who to give the data to, but I think Google is really high on the list because they've been around for so long, and I think that they've already won a lot of trust.
Paris: I'd say Google, Apple, who is yet to really enter the game in a serious way, but Apple's got that level of trust, I think. But ChatGPT, and you see these are still relatively new companies, and I can understand why people are still hesitant to hand over this data
Christian: Yeah, and especially, I mean, you know, a lot of people I know use ChatGPT or Claude to build on, you know, services or platforms that have recently entered the game as like a quick way to spin up a website or an app or And you know, you don't know a whole lot about them. And so I always just because I'm more familiar with it, I always default to...
Christian: And, you know, everything in our company runs on Google Workspace, so we can, you know, we can integrate this with every other part of the, you know, every other part of our, architecture. You know,
Paris: And yeah, and BigQuery makes a lot of sense, too. We are always internally at our agency, we have a BigQuery instance and everything goes there. And, we're all in on Google's entire stack, Google Workspace and Cloud, BigQuery and the rest. Are you able to successfully get this implementation done with clients to have them really embrace the idea that we're gonna create this central knowledge base, this hub, and we're gonna voluntarily flow our chats in there, emails, meeting transcripts?
Paris: 'Cause what about confidentiality, privacy, even Slacks between, you know, one-to-one DMs? You're starting to get into some tricky territory there, I guess, with people
Christian: Absolutely. Yeah, for sure. And I dedicated a whole slide to this question, of basically, like governance and security, during my talk, and y-you're absolutely right. It's a really important point. And yeah, for sure, getting customers on board with this is obviously like a hands up, like not we're not doing this with every single company. I'd say it's the Of, the companies that are more open to experimentation and who are, quicker to adopt kind of AI methodology, now it works, and it works well.
Christian: But There's a certain level of education and governance that has to go with that. But essentially it comes down to just, having a really clear picture of, what you want to, what you want to share and ingest, and what you don't. And it's the same in our company. Like we, you know, there's very strict, controls around what channels might be looked at.
Christian: So, you know, no private conversations, obviously. We only have one or two channels in, ClickUp, that we like automatically crawl, and that's really just the, that industry news one that I mentioned. We don't, you know, we don't connect it to any of our client channels or anything like that. And so from our customers, it's really o very often it's a case of showing them what we've done.
Christian: So, you know, saying like, "Really useful. If you talk about this kind of stuff, it's useful to have a place where you can do so that you don't mind being public. So create a Slack channel for it, and drop things Have discussions about them." and other than that, we've just tried to have, you know, a simple enough interface. So as I mentioned, one of the, I think, really useful ways to do this is that you can set up an email address that people can just email stuff to, Into this database.
Christian: So if Or if there is a transcript that you want to include, you just forward the transcript rather than relying on, you know, more automated channels, which, yeah, of course, you know, you wanna have as much control as possible. And obviously, that limits the amount of information that's gonna come in because, you know, people are being more circumspect, but that's a good thing.
Christian: And I would say it probably makes for a better database because you're, you know, you're really curating what goes into it, and so you can choose the LLM really understands business and your direction and your kind of thinking about things, know, in a really specific way.
Paris: Yeah. Yeah. I think that's, yeah, that's definitely the right way to do it. DMs, any one-to-one messages, I think always need to be excluded. But when there's a project space that could be... We don't use Slack, but we use Google Chat, which is the Slack alternative. Have you tried? Yeah. It isn't as good as Slack, but we made that decision years ago and, Yeah, it does lack in some of the functionality of Slack.
Paris: And b- and by the way, every client is on Slack, so we, by default, we have to use Slack, and we get into the Slack channels and spaces of the clients. But we have a project space set up in Google Chat for every project, and that's really the place where everything that is discussed in there should be... People should know, I mean, this is the stuff that we will We really need to use this as context.
Paris: And so if... Yeah, if something private needs to be said to someone, you do that in a DM, and that's gonna be safe and off limits from any kinda, any kind of knowledge base. But the project spaces are really gonna be made publicly av- Well, not publicly available, but that's gonna go into the knowledge base. And it is a mindset shift, I think, for a lot of companies to adopt that mindset that we need to start capturing everything that's said in meetings and in all these, this chat activity and by the way, also Google Drive too.
Paris: All of the, all those decks and documents and spreadsheets and all that, It's a wealth of knowledge about the company
Christian: Yeah. I think there's two important things to remember. You know, whether you're taking like, you know, a very like structured and organized and kind of like technical and methodical approach like we have for ourselves, where there's like automations going and pulling in stuff and or automatically organizing content, or whether you're doing it kind of more manually. You know, as again, as long as you're using a surface like BigQuery, for example, it's there, but it's safe.
Christian: You know, there's no public access, and nothing's happening with it. And equally as importantly, you know, we don't generate, content. We don't we use human writers, and the most that we might do, using this, you know, using a database like this is creating a brief. Because you can pull in like, oh, on this date, this person, you know, talked about this, and on this date, the founder, had
Paris: Under this,
Christian: And they reacted to this post. And you can, you know, you can really easily pull in five months of conversations, documentation to everything and say, "Okay, what does the company think about this one topic?" And then, you know, that makes for a really great blog article. But it's still Writing it, and the you know, as long as you're able to trace back everything to an original source, which again is something that we've, you know, we've set this up so that any quote, any insight, anything, it's basically like, you know, in Claude and ChatGPT, how it cites the online sources that it's using, and you can click on them and see them.
Christian: This is gonna cite back and say, "Oh, this was a Slack conversation." You know, within the editorial process, you can immediately see, oh, well, something we said about, you know, with a client. We definitely don't want it, so we'll not use that. Or, oh, we definitely want to, you know, we definitely want to, include this. And then the other thing that companies should remember is, you know, unless they've very specifically opted out, this is all happening anyway. Slack now has a brain. ChatGPT, ClickUp now has a brain.
Paris: Yeah, keep in mind Slack is owned by Salesforce. Yeah, S-Slack is owned by Salesforce. Th-they're, what They have an agent. What are they called? The, I mean, they have a huge agent, AI agent platform. Agentforce, is it? I think it's called Agentforce. I mean, it's probably one of the
Christian: Private con I mean, fingers crossed, hopefully not private conversations. Like, I mean, that would I think a,
Paris: Yeah, that would be crossing the line. Yeah
Christian: Of companies, you know. But, nonetheless, I mean, these models, you know, they're not on-machine models. So, like, when you ask, Slack AI or ClickUp AI or even Gemini, you know, and Google a question, it's essentially doing the same thing as this. A-another person I spoke to kind of equated this to NotebookLM. And it's true. This is it's similar to NotebookLM.
Christian: It's just that we've set up, you know, kind of unique ways to ingest the thought leadership, like ingest the experience and expertise. And then we've made it a little bit more useful than NotebookLM. But, You know, it's all happening already in the background, so you might as well develop a system that allows you to control it work hard for you know? You Something out of it. You know, Getting a lot out of your data. You should get a lot out of it too
Paris: Yeah. Yeah, absolutely. And, funny you mentioned Google NotebookLM. I was I think they now call it Gemini Notebook. I'm a huge fan of that product. It's I think it was one of the under-the-radar AI products, and one of the That was really one of the inspirations for our internal tool that we built called GEOforge, which is that all content really needs to be grounded in some sort of a source material or a knowledge base.
Paris: And with NotebookLM, in order to generate any kind of a content, you need to put in sources, and you need to create that context or that knowledge base. But you can't go to it like you could go to Claude and just say, "Write me an article about topic XYZ." And, because Gemini Notebook, it needs that source material so it can reference it and you can use that.
Paris: I guess it's using some kind of a RAG, a RAG-based pipeline. But that was one of our inspirations because we have, our platform has a knowledge base that grounds all the content, and you literally can't even create anyth-anything until you've got a knowledge base with some documents that it can reference. Yeah. So it by definition avoids any kind of AI slop
Christian: There's a new tool that, Google launched, yeah, oh, a year ago, actually. Well, I only learned about it recently, called, File Search Tool, which is it's essentially a managed, RAG, system. So essentially like NotebookLM, like Gemini Notebook, except that it runs through API. So I've actually been looking at whether this would be a better solution, for example, than BigQuery.
Christian: But I like that with BigQuery, you know, we can add metadata, so we can tag content, with, you know, the date it was created, who was involved, what, theme does it fall under, all of these kinds of things.
Paris: Uh-huh. Yeah, that's interesting. So if I heard you correct, you're still actually drafting... Even the first draft is written by human writers still at Eleven? Yeah.
Christian: Yeah, it is. And one of the main reasons why, again, is because of this idea of, like sourced expertise, and we really believe, and our experience has shown us that the best content comes from people who understand what they're writing about, and who can, you Can write about it with nuance. But at the same time, you know, they need to be able to understand the business that they're, you know, that we're working with and understand the data that they're looking at.
Christian: But beyond that, for a lot of reasons, including editorial reasons, you know, we and most of our customers, you know, you can tell. But you know, there's still a lot of studies that show that human readers respond better human-written content, and that human-written content tends to, delights us in different ways. It surprises us in different ways. It tends to be less linear and straight, and like when you graph out the use of different kind of writing methods, ChatGPT and Claude, they all tend to group to one side and all do the same thing, and then there's this vast, unexplored kind of literary and editorial, surface over here.
Christian: And, you know, what we've seen, is that the engagement tends to be higher and conversion tends to be higher when people feel like they're reading, know, something from another human individual like them.
Paris: Yeah, I think that is true. And actually I, at brightonSEO I met AirOps' Sam Page, and we were in a, o- one of these, round table sessions together, and he made that point actually, in the group that they have data that proves that the human-generated content, 'cause they actually ran some testing on this, that human-generated content per- even performed better in LLMs as compared to AI-generated content.
Paris: And so I do think even LLMs when they're crawling it, they may be trying to detect the degree of human effort that went in. Whether that's, I mean, human effort meaning that could be whether it produced the original draft or whether w- there was significant human editorial touch, on top of it. 'Cause the approach that we take, and within GEOforge is that the first draft is AI-generated, and I do think that that's okay as long as it is grounded with, it's grounded in that knowledge base, the proprietary knowledge base from the brand.
Paris: If you have AI draft content that goes out and just does web research and it's just purely derivative, That adds, that really adds no value. I mean, that used to be a popular SEO tactic. You probably remember the term the skyscraper technique
Christian: Yeah, of
Paris: When you Yeah, you're just basically trying to pull the best parts of all the top-ranking pages into one super page and that used to work. I mean, we used to do that too back in the day. But now, of course, you would never do that now because, that's AI clearly can synthesize better than, anybody To do that manually. But it's the truffles again.
Paris: I mean, you've got to unearth those truffles, serve them to AI, and then I do think our approach is that first draft just to expediate the process, the draft is generated by AI grounded in the knowledge base, and then the human comes in immediately and, fact checks everything, and we have fact checking both, on claims made that are connected to the data, to the knowledge base, And then other facts that may come from other... There is a degree of web research that also folds into the piece
Christian: Yeah, of course.
Paris: And that needs to get fact checked.
Christian: Everything in context for sure. You know, as we've said, the most important thing is just, originality and information gain as you've said. And so where that comes in is probably less important than the fact that it comes in at all, you know? An AI-assisted, article that, is the fruit of, collection, analysis, you know, all of these really unique things to a company be always be better than a fully human written piece that's simply a synthesis, you know, or, you know, just a, you know, kind of research or skyscraper method, To be.
Christian: You know, their content is really good and really interesting, primarily because it's so deeply rooted in experiments that they run and analysis that they do, and then of course, humans that shape the narrative and sit down with the data and think, "Okay, what is this saying about the industry? What is this saying about the world? What does this say about, you know, about us?" And then if it happens to be a machine that puts that into, you know, who cares who's typing, right?
Christian: Thinking and the analysis and the intuition that comes from humans that I think is the most important. And then, you know, the actual hands-on keyboards less so.
Paris: Yeah. Yeah. Christian, do you have an opinion about the right amount of content, the content volume? Because now AI clearly empowers content marketers to do a lot more content. Let's assume that we're like an Ahrefs, or we have a wealth of proprietary brand knowledge and truffles that we can access, and the question then becomes, well, how much content should we produce, and w- how much is too much?
Paris: Is it more than one piece per day? Or, I know it certainly would depend on the brand and the industry, but do you have an opinion on where is the line for volume that we probably should approach but not cross?
Christian: So, I would say that my the line for me is not, necessarily an exact amount of content. It's more a question of what's the most amount of content that you can reliably produce before it becomes derivative and repetitive? So if you're Ahrefs, that's probably a lot. You It's a massive company. They have of data points from, you know, from their own index, from their millions of users.
Christian: So like any enterprise, you know, level company dealing with a lot of data, that upper limit of, okay, how much can we produce here before we just start repeating ourselves, or it's just, you know, being the same stuff over and over, is probably gonna be pretty high. For a company that's, you know, smaller or maybe not necessarily so, data-driven, or is just focused on other things, then I would say it's probably going to be less.
Christian: And for me, the more important, thing is to, Yeah, just to not try to produce content for the sake of producing content, because then you're producing content that doesn't have substance, Really identifying where that line is. So for one company, you know, it might be quite a lot a month. Now I'm a big fan of a lot into key assets like e-books and white papers and these kinds of things, and then building up content around them that supports them, so blog articles, LinkedIn posts, newsletters, all of these things.
Christian: And of course, if you're taking that view, then there's quite a lot of content that you can create around one single kind of fundamental, pillar or idea. And, HubSpot, I think it's HubSpot, recently did a presentation, somewhere, probably some HubSpot event, that my- one of my colleagues shared with me, about this new kind of content cycle that they've created. I didn't agree with everything in there, but there
Paris: Oh, yeah. That called the Loop, by the way? Do they call that the Loop?
Christian: Yeah, yeah, the Loop.
Paris: Yeah, that's their new mental model. Yeah, they're pushing that a lot.
Christian: Yeah, yeah, their Loop. And I didn't... You know, I think there was stuff that I agreed in there and stuff that I didn't, but one of the ideas that I really liked from them was this idea of being, relevant and relentless. So you have to be, relevant in that, you know, you have to be E-E-A-T, you have to be unique and everything, but you also have to be relentless in demonstrating that to the world.
Christian: And so I think that there's probably enough opportunity for any company Like there's so many surfaces, there's so many channels, there's so many networks, there's so many ways to talk about the same things to different audiences. Reddit, Quora, like, you know, Things, that, you know, there really is probably more content that one could create than any company, because of budgets and others constraints, is going to create.
Christian: The line again for me is just about, diversity of the content and making sure that every piece of content has a real purpose. Does any company need to do 50 blog articles per month? Probably not. Don't do 50 blog articles. Do three e-books, 10 blog articles, LinkedIn posts, get on Reddit, you know, build up this kind of machine around it, and Probably the better use I would say. So that can still be a lot of content, in the variety.
Paris: Yeah, yeah. When you consider all the repurposed assets and what you described, I think this was originally a HubSpot concept too, but the pillar cluster model, Think that's kind of how you've described it, where you have some pillar pieces. It could be e-book, white paper, or it could be just a really in-depth guide to something. And then you might do one of those a month, and then around those you can have all the supporting content and repurposed assets.
Paris: And I think being relentless is a ... It does mean repeating things a lot too. I mean, if you have your core f- your core beliefs and your philosophy and your differentiators do need to be repeated to the market a lot actually. But then you don't want, you don't wanna just keep producing, you know, identical types of content. So if you've ...
Paris: I think where you've crossed the line, as you said, is when it really just becomes purely repetitive and you're not saying the same thing in a new way for a new audience un- or through a new format or through a new, channel. But it's just, purely repetitive, that's when you should probably just pull back at that point. You know, that's enough
Christian: Yeah. I mean, there's always new ways to talk about the same thing. And a company should stay true to its values. And, you know, some of the... You know, when I was starting out very, early on, one of the best pieces of advice that I, remember getting was, you know, pick three or four things that are really important to you and just create lots of content around that.
Christian: And, you know, the way that you do that is you look at it from different angles, you look at it from different perspectives, you tie it to different things that are going on now in the market. So for example, Eleven has always been about topic expertise. Expertise has always been this thing that's like been like very deeply rooted in our DNA.
Christian: So now we talk about... You know, five years ago we were talking about expertise in SEO. Now we're talking about expertise in GEO. Tomorrow we'll be talking about expertise in who knows what. So there's, you know, kind of infinite opportunities, to talk about the things that you feel are important as a company, and still making it, information gain, still making it unique and original and interesting to buyers.
Paris: Yeah, I totally agree. Let's pivot over to the SEO versus GEO, I guess the debate, so to speak. A lot of GEO tactically looks similar to SEO, but then you could also argue that GEO is something that's bigger because it's more about brand positioning. But where do you fall in this debate? How do you see G- SEO and GEO?
Paris: Is it a Venn diagram where, you know, one of the bubbles is bigger? Or are they How much overlap is there? Is it the evolution of SEO, or is it the emergence of something really different?
Christian: Yeah, yeah. So, very good question, obviously. And, just as you were asking it, I was thinking about, you know, how different people might respond to this question. Because I think if we're coming at it from the perspective that I would typically come at it, which is what are the things that a company needs to do to be successful in SEO or in GEO?
Christian: For me, there's quite a lot of overlap. You know, other than the, you know, specifically really technical bits like crawlability and indexability and these kinds of things, I would say that, what we feel works really well anyway for SEO is also what works really well for GEO, which again is this idea of, demonstrating E-E-A-T, demonstrating originality, uniqueness, high effort content.
Christian: You know, Google just recently updated their helpful content, guidelines, to include effort, originality, talent or skill, and accuracy. It's interesting that they put accuracy in there. So all of these things are, you know, they're going to be beneficial to SEO, and they're gonna be beneficial to GEO, and they're gonna be beneficial also in the most important surface that you're trying to influence, which is of course the brains of your consumers, you know, the minds of your main audience.
Christian: And so I think all of the non-gimmicky, non-hacky stuff, you know, is really fundamental to, both of them. I think probably where it differs is the degree to which it's going to influence it. Digital PR, for example. So as I mentioned at the start, Eleven A digital PR arm, that we're very excited with. We've got a number of clients that have signed on.
Christian: It's very cool to be using this, and we've brought in some, experts to run it. Would say that that's very beneficial for SEO. I would say it's critical for GEO, Because of this idea of, off-page, you know, the importance of off-page content in, shaping a narrative, on, GEO. And of course, that's because GEO presents, information, in a... You know, it tries to do it in the way that a human would.
Christian: SEO, predominantly was just links. And so, you know, digital PR and PR, traditional PR were ways to, you know, manipu- not manipulate, were ways to improve, You
Paris: It, yeah, it was a link building tactic essentially. What PR in an SEO world, yeah
Christian: Yeah, exactly. It's a link-building tactic, and we knew that that was, a way to demonstrate to Google that, hey, we have ties to important people. We're, you know, we're trustworthy. We're an authoritative, website. But in LLM models, you know, because they're presenting information in a conversational way and LLMs permit themselves to have an opinion or to express an opinion, in a very human way, "Oh, you should be careful of," or, "This company might not be u- for you because they don't have this feature or this feature," or, "This feature is very highly noted," but, you know, like, there's a very, like, there's an opinion, you know, there.
Christian: Now, whether there should be an opinion or not is another conversation. But there's a what people read to be kind of, you know, an AI opinion. And digital PR influences that. It influences directly how an LLM talks about you. So it's not just a technique for, you know, kind of improving rank or even in an LLM improving how often you appear.
Christian: It's a tactic for shaping the narrative when you appear. I think that's true for almost all of the tools that were in our SEO tool belt, and in our, you know, brand marketing and product marketing tool belt. Now they all influence, GEO. And it's like we were, you know, you and I had a conversation before where we were saying that GEO has kind of started to pull together these different marketing departments, product, brand, SEO, et cetera, because now all of them influence the way that LLMs talk about you
Paris: Yeah. Yeah, I think it's a great development for the industry that GEO now has a seat at the big boy table, so to speak, where
Christian: It,
Paris: Probably the CMO is paying more attention than they ever did to SEO, and the CEO even probably is paying attention. And yeah, I think it's about the
Christian: Are all over it.
Paris: It's kind of the equivalent of when the CEO would go to Google and Google their own name and to see what comes up, or, I don't like this. We need to change this." But, of course, yeah, now it's a, it's a much more complex 'cause the CEO doesn't just see a bunch of links, but they're really seeing a story about themselves or about their company.
Paris: And, I think, yeah, in an SEO world, the authority primarily flowed through links, but now in a GEO first world, authority is something much broader. It f- it flows through, general association with other brands, and that, those associations can be links, but they can also be just much broader. They could be mentioned and the collaboration with brands is more apparent.
Paris: And I think it really comes down to like a co- a co-occurrence type of, the math behind it. You know, how often are we seeing this brand mentioned alongside other leaders in the space? What are they saying about it? And then can we corroborate all this evidence into a fairly consistent story about the facts and about these entities and facts about this brand?
Paris: And if so, then we will trust that brand enough to believe what they've published on their own website, which is, of course, anybody can publish anything. Yeah
Christian: And have you noticed that, there's been a couple people talking about this, but, ChatGPT and Claude have started to call out when information is coming from a website of its own. So like, if you now ask Claude for a roundup of the top, you know, five, you know, kind of whatever software, it will list it and then it will say, "However, many of these recommendations come from the producer, the vendors themselves and so should be taken with a grain of salt.
Christian: Here are some more independent, things that you can look at." And you know, that's really interesting and obviously coincides with the fact that, Google as well, has been, if not de-indexing, then certainly demoting, self-serving, or self-promoting, you know, self-promoting listicles, which, you know, really had their heyday in the last year and a half
Paris: Oh yeah, the listicles, they w- they won't die. You know what's interesting what you just mentioned because I, the primary way that I produce this show is that I get a daily intelligence briefing from ChatGPT with a bunch of stories about I get 10 stories a day, and ChatGPT in almost every story there's a the last line of that snippet is some sort of a disclaimer of what you just mentioned where it says, "Now this, now these, I don't know, this study should be not taken as representative of the whole industry because this is only coming from, AirOps.
Paris: This is a limited study into this field," or something like that. But it there's I think that's also a similar thing that AI is saying, "Okay, this is what this brand reported in its own website, but we can't necessarily corroborate this across a lot of other sources." But nonetheless, it still makes it into the brief of the, you know, what should be the most interesting 10 stories of the day for GEO
Christian: For sure. Absolutely. It still makes it in there, and I mean, you know, immediacy bias and first sight bias says that whatever people see first, they're gonna put more stock in. So it pays to be in there, even if it is with a caveat. It's better to be mentioned with a caveat than At
Paris: You're still mentioned. Yeah.
Christian: 'cause Yeah. As long as you're ment- again, as long as the mention is a good one. You know, we've seen, examples, recently, not, with our customers, but, other people's research that I've seen, calling out the fact that, you know, it's not enough just to count your certainly not enough to count your citations. I mean, who cares about citations? But it's not enough to count your mentions.
Christian: You have to look at the sentiment around them, because You know, Claude or ChatGPT will say, you know, they'll mention a product, but they'll mention it as it, they'll mention it saying, "This one probably isn't for you because it Feature." Now, maybe you have that feature and it's Properly displayed or, you know, you're not talking about it enough. You know, ChatGPT, again, it permits itself to kind of
Paris: Absolutely. Sentiment, sentiment is huge.
Christian: It.
Paris: Yeah. And sentiment didn't really exist in SEO because SEO is just g- with the just giving you the links. So you can interpret the sentiment any way you want. Here's the data, and you do the research. Sentiment is very important. What, I mean, what if the user's prompt was, "Which companies in this space should I avoid working with?" You know, and you don't wanna be... You don't want that mentioned. Yeah
Christian: Don't want that. Yeah, exactly. You don't want that mentioned, yeah. It's interesting because, you know, especially like, you know, with SEO, people would land on your site, and you get so... It's so much easier to control, you know, what they see. It's so much easier to control the narrative and
Paris: Yeah, and the whole journey.
Christian: You know, for better or for worse, again, because and Anthropic and Google have created these, LLMs to express an opinion and to speak to you as if it was a reasonable, rational, intelligent individual who was making decisions. It's not, obviously. It's not. Algorithm, you know? It's just very sophisticated, you know, next not even next most likely word, but word that's most likely to please the end user.
Christian: So we But nonetheless, our poor, weak human psychologies are incapable of reading it with as much, I think, know, critical, regard as we should. And Take what it says. We're very keen to be told a decision. I think, you know, We just wanna know what's the right one for me? Was, It's my, husband's birthday in a couple days, and I was, wanted to do a present for him.
Christian: So, I mean, this will come out afterwards anyway, so I don't have to worry. But one of his, favorite, comic books is, called, Méta-Barons. Les Méta-Barons, it's French. The Méta-Barons. And so I took, I found, like, three original, cells, and I'm getting them, printed and framed. And I was speaking to the guy, and he was like: "Do you want this or do you want this? Do you want this or do you want this?" I was like: "I don't know." Like: "I don't know. Just tell me. Just tell me what
Paris: Yeah, yeah. People definitely want shortcuts to the tough decisions in life
Christian: Gonna look good. Just, like, make the decision for me." And I think And, you know, I see it with myself using LLMs, and every time I see myself like: "No, Christian, like, apply brain to task.
Paris: Yeah. Well, you've got that psychology degree, so you have a better, you have better discipline from your training than most people
Christian: Told what's the answer, and these companies have given us something that, you know, very often will Or at least before. Now they're doing a, now they're doing a better job of saying, you should verify." And it's more than At the bottom saying, "Claude can sometimes make mistakes." You know, Claude itself is actually saying, I You know, you need to make your own decision here.
Christian: Here's what I can tell you, You should look into these." and that is getting better. But at the end of the day, you know, it has a lot of influence on human psychology. And so coming back Business side of it, then, you know, yeah, you need to make sure that you're being presented faithfully and appropriately, you know, by these platforms, because it's no longer a space where you can fully control the narrative and you can choose what colors the user sees and where the call to action is and all of these things.
Christian: Like, your website was a very, you know, it was this little area that you could control. That's not where most people are interacting with companies now.
Paris: Yeah. Yeah. That whole journey has really collapsed in a sense into chat. And when people do show up, in your store, they're showing up with an opinion and with some motivation. And most likely they're showing up with stronger motivation to either to buy or to convert or, Take that next step
Christian: Higher for
Paris: If they're there.
Christian: From They're many points higher, you know, like, 3% from Google and 13%, for referrals from ChatGPT. So they've When they do come through, they've formed an opinion. But, you know, they have an opinion, and that opinion might be borrowed from an LLM. So what's the LLM's opinion of you?
Paris: Yeah, absolutely. And that trust with, as we're entering this new chapter of AI, personal AI agents, for lack of a better word,
Christian: There's so many announcements.
Paris: But when those personal...
Christian: There's
Paris: Yeah. It's hard. Yes. I'm trying to keep up with it all on the GEO Show, but it is, it's like drinking from the fire hose. But in any case, the, these agents that are now becoming more proactive, and they're actually... They're not just waiting for you to come and enter a prompt and try to do some task, but they're proactively saying, "Hey, heads up, you know, we there's an email that you need to respond to that's now two days old," or whatever.
Paris: That just builds more and more trust between the user and their personal agent, and then when that personal agent makes a recommendation, it's even more so gonna be like, "All right. My Grok Bot tells me this is the best choice. I'm going for it. I'm gonna offload my cognitive load of this hard decision
Christian: Absolutely. And again, like, you
Paris: It."
Christian: It. Exactly. That's what it is. Outsource. And again, I have to say, you know, like for better or for worse, because, you know, I think in a lot of cases it's for worse. And not I mean like for us as a species, you know, for us as a society, outsourcing, you know, critical thought, It's
Paris: That's exactly what I was gonna say
Christian: Me think of, the, that line in "The Matrix," where, Agent Smith says, to Neo, you know, and I say your world because, you know, once we started to think for you, it really stopped being your world, and it was our world." And it's, you know, What is our world right now? What kind of world are we And I'm, you know, I'm not an alarmist about, LLMs. But you know, I see myself ceding more and more critical thought to, you know, to LLMs 'cause Is there, you know,
Paris: It is. In the relentless pursuit of productivity, yeah. I, and I think, yeah, we're... I, and I feel it with myself. I am striving so hard for more productivity with these tools, and it makes it so easy that, okay, I can make I need to make decisions faster, so I'm gonna outsource this. Yeah, I don't see this trend reversing for...
Paris: Yeah, as you said, for better or worse. I mean, we're barreling into this new world and, there's implications with all of society, education, you know, critical thinking. We can't, we can't lose that. That is a real innate human trait. So, well, yeah, this has been great, Christian. I, we,
Christian: This has been really interesting. Yeah, I
Paris: Yeah, yeah. Thank you for the time and, it's great work that you all are doing. Just to wrap up, where can people find you online? How can they connect?
Christian: Yeah, for sure. So you can find, the company at 11content.com. That's 11 like the number, 11content.com. And you can find me on LinkedIn. I'm just Christian Rigg, R-I-G-G.
Paris: Great. All right, Christian, thanks for being with us today
Christian: My pleasure. Thanks so much for having me, Paris. Bye
Paris: All right
