GEO Expert Adrian Dahlin on Measuring AI Visibility

Adrian Dahlin is a GEO expert, data scientist, and founder of Search to Sale. In this live audit with Dane Frederiksen, he shows how to measure AI search visibility and use the findings to plan useful content.

For B2B marketing teams, the point is to see where buyers can find you and where they cannot. Digital Accomplice uses that evidence to build video plans around real buyer questions. Start with the questions your sales team already hears, then check what the answers currently show.

Key Takeaways

• Adrian Dahlin separates brand mentions, URL retrievals, and citations before interpreting visibility data.
• Digital Accomplice tests prompts around buyer roles, services, industries, and geography.
• YouTube led domain retrievals in this audit's selected prompt set, not across every industry.
• A vendor comparison surfaced as a content idea, not a proven revenue result.
• Repeated observations help guide experiments; they do not guarantee recommendations or new customers.

This fits inside a bigger picture. If you want the full view first, here's the bigger picture.

Why is one ChatGPT search a weak baseline?

One personal search does not tell you how a whole market sees your business. Your conversation history and the context you provide can change the answer. Adrian Dahlin, founder of Search to Sale, explains why he uses a structured prompt set instead.

The starting point is the buyer's situation. A broad question about project management software differs from a question that names a role, company type, past tools, and a specific need. The same applies to choosing a video partner.

In this audit, Dane and Adrian examine prompts about video strategy, production, AI search visibility, and the kinds of companies Digital Accomplice serves. Those prompts are research hypotheses. They are not a transcript of every question real buyers ask.

Adrian's approach is to run those prompts repeatedly across different systems. That creates a record the team can compare over time. It is more useful for planning than checking a personal chatbot once and treating the result as a market ranking.

The limitation matters. A measurement tool still samples selected questions and conditions. It cannot recreate every user's history or every possible buying journey.

What do mentions, retrievals, and citations mean?

A brand mention and a cited source answer different questions. The audit keeps them separate so a promising number does not hide a content gap.

A mention means Digital Accomplice appears in the generated answer. A retrieved URL appears as a linked source. In Adrian's explanation of this tool, one retrieved URL can receive multiple citations within the same answer.

That means a brand can appear in an answer without being the source behind it. It can also have useful content cited without receiving the kind of recommendation a buyer needs to choose a vendor.

In one selected prompt, the dashboard showed a 28% mention rate across the audit's tracked answers and review window. That is a result for that prompt and sample. It is not Digital Accomplice's share of all AI answers, a conversion rate, or a forecast of future leads.

Dane keeps bringing the discussion back to the business question: does being mentioned help someone consider you? The audit is a way to investigate that question. It does not settle revenue attribution on its own.

How do you choose prompts that fit your buyers?

Build prompts around the services you want to sell and the people who could buy them. Then check that those questions have a basis in customer conversations or search research.

In the demonstration, Adrian combines several dimensions:

• The buyer's role, such as a VP of marketing.
• The company type, such as B2B SaaS or cybersecurity.
• The service, including video strategy and production.
• The goal, such as answering buyer questions or improving AI visibility.
• Geography when local work is relevant.

One prompt describes a narrow fit for Digital Accomplice. Adrian explains that this is intentional: first establish whether the business appears for a description of the specific thing it does.

Dane adds another source of useful language: sales transcripts. Buyers may recognize the problem before they know that video strategy is a possible solution. Their questions can help the team avoid writing only for people who already know the service name.

This is a starting map, not a finished content calendar. A prompt still needs a helpful answer, a sensible priority, and editorial judgment before it becomes a video.

How does the audit help decide whether to make video?

Look at the sources retrieved for your selected questions before committing to a format. A platform's overall popularity does not prove it is the right place for your particular buyer problem.

In this audit's tracked prompt set, YouTube was the most retrieved domain. That gave Dane and Adrian a reason to inspect the actual videos behind those results. It did not prove that any video would work or that video should replace website content.

For one vendor-selection prompt, written lists of production companies appeared among the cited sources. Elsewhere, a video comparing agencies surfaced as an example worth studying.

The useful next question is what those assets help the buyer decide. Are they defining a service, comparing alternatives, explaining a process, or demonstrating expertise? That distinction is more useful than copying the format without understanding its purpose.

As Adrian puts it: "It's much more helpful for us to look at what type of content is influencing AI answers in your space."

Dane's contribution is the video judgment. The audit identifies something worth examining. He then evaluates how to make an honest, clear answer that helps the intended buyer.

What should the next content experiment look like?

Choose a question that matters commercially and where the audit suggests a visibility gap. Then make something that helps the buyer evaluate their options.

Dane and Adrian consider a comparison of video agencies, consultants, or strategists. The idea is to explain when Digital Accomplice fits and when another option fits better. Dane explicitly says he wants to equip buyers to make a good decision rather than simply praise his own business.

That is an experiment, not a result. The interview does not establish that a comparison video will win a citation, produce a specific number of views, or close a new deal.

A practical starting point is to choose one priority prompt, inspect its sources, document the unanswered buyer need, and make the most useful response you can. Keep the original prompt and research conditions so later observations have context.

Track citations separately from visits, inquiries, and sales conversations. A changed answer is evidence about that observation. A qualified inquiry is evidence about a different part of the buying process. Both can matter without being interchangeable.

Doing the thing is the easy part. The hard part is knowing where you're already winning and where you're invisible. That's exactly what our free AI Visibility Snapshot shows you, before you spend a dollar.

Frequently Asked Questions

How can I measure my company's AI search visibility?

Test a defined set of buyer prompts repeatedly and record when your brand appears and when your sources are cited. Keep the systems, dates, and prompt wording with the results.

Does a brand mention mean my website was cited?

No. Adrian distinguishes a brand appearing in answer text from a linked URL being retrieved. A citation count is another measure in the tool's reporting.

Should every B2B company publish more YouTube videos?

This audit does not support that conclusion. Inspect the sources behind the questions your buyers ask, then choose a format that serves their decision.

Can this audit predict how many deals I will close?

No. When Dane asks about the effort required to close a new deal, Adrian says he cannot answer that from this audit. Visibility and revenue need separate evidence.

What should I do before planning another video?

Choose a buyer question, check its current sources, and identify what a better answer would help someone decide. If you need help turning that into a B2B video plan, talk with Dane.

Most teams don't get this right on the first try. Here's how we approached it for a real team. Curious what it'd look like for you? Let's talk it through.

Full Interview Transcript

Source: Riverside transcript, trimmed to the edited interview. Minor wording and timing differences may remain.

Dane: Hello, everybody. my name is Dane Frederiksen. I'm joined today by Adrian Dahlin. We are going to get into AI visibility and what you can do about it. for a lot of companies it's a top priority, and they're thinking, how do we show up in AI search? how do we know that for sure? And then what are we gonna do to engineer more visibility like this? And then

Adrian: you

Mm-hmm.

Dane: Ideally getting cited and getting recommended is the real goal to drive pipeline. But obviously if they can't find you in the first place, they're not gonna be doing business with you. So, Adrian, do you wanna kinda talk about the tool that we're using and why this is the right tool to measure AI visibility?

Adrian: me.

Adrian: Yeah. Okay. So if your goal is to be present in AI answers, we need to have some way to measure if that's happening today. And you of course could go into chat, GPT or Google and look at an AI overview and test out a prompt and be like, if you sell project management software, you can go ask chat GPT.

What are the best project management software solutions? But you're always going to get personalized results when you use these yourself. Plus, you really want to have a structured way of testing a variety of different prompts that people might be using versus just best tool for X.

And this starts to get at the, one of the things that's different about AI versus Google and search engines that we've been working with for a quarter century. And that's that we are now in the era of the long tail. The way people use AI, it's different from how they've been using search engines. We give much longer questions to AI. We give it more context. It knows more about us.

of all of our history of interacting with it. And in every given prompt, we tend to give more specificity in our question. We tend to describe our situation, our need, our problem. And that means that we're getting two longer tail answers, where it's not best project management software. It's here's my role at what kind of company, here's the kind of projects we work on.

I've used Basecamp and I was frustrated about that. And I love this. What should I, what project management tools should I consider? And given that it's a long tail, that means there's many more variations to be considering effectively more keywords to be targeting. And it's super helpful to have a system that can track a lot of different AI prompts and run them.

Adrian: on external servers that don't have all of your personalized context and give you structured data to see when someone asks this question, is my brand mentioned in the generated answer? Is my content cited? My website content, my social media content, is it being cited among the sources? How often is my content being used as a resource by the AI?

generates its answer. And then as you start to then develop content strategies based on what you see as influencing AI, you start putting more content out there, what changes, what new content has an impact on AI answers. That's why we use a tool like this. This one is peakpec.ai. There are several others out there. Comes through profound.

Dane: Right. So essentially, you know, what I'm hearing is your CEO wants AI visibility. You use a tool like this to find out where the holes are in your content plan. And then you can go make that content and then find out later which of those holes have been effectively plugged and then what still needs more work. And then you're starting to approach a state of like content bingo where you've got all the squares covered, and then ideally then you're the most visible and you're the one that gets recommended. Does that sound like the right plan?

Adrian: Yeah. Yeah.

Adrian: Yeah. And let's like sanity check. Is this worth it to at the start? ChatchGBT is like the sixth or seventh most used website in the world. It is still way far behind Google in terms of daily usage, but a large and increasing share of Google searches feature in AI overview.

So it's pretty clearly helpful for a brand to be showing up in AI Answers as long as you're optimizing for Google's AI first, chat GPT second, and everything else after that.

Dane: So let me just kind of make sure I'm understanding. Like we're saying that this is important, but this it seems mission critical to me, because like if you're not visible in an AI search world, you're you're basically just your business is dying unless you're like you've got enough referrals to keep you going.

Adrian: Yeah, yeah, there's different ways to get business obviously, but yeah, it may like it's already having an impact now. so Google's traffic from Google out to other websites. In other words, the share of Google searches that results in clicking one of the blue links has gone down a lot, maybe ballpark 30 % in the last couple of years.

That's because of all the engagement with AI, because AI, the overview is answering a lot of the questions that people went to Google for and there's no click through happening. But there is brand visibility happening, whether or not there's a click through. That's what's changed already today. Plus, let's imagine in the future as AI, as usage of AI as an interface and as an intermediary.

between us and all of the world's information increases. AI is only going to be more of like a curator of the information we're consuming. That includes within Google, that includes dedicated chat bots like ChatGPT and Cloud. It includes other interfaces like there might be a lot more voice in the future, like these devices that Apple and OpenAI are developing, you know, in-home devices. It might be there's lots of voice interface.

which means we need to figure out now how to be found by AI so that we can service not only in aerial reviews and chat GPT today, but the kind of AI interfaces of the future as well.

Dane: Yeah, future proofing here. So so what we're gonna get into here is we're gonna actually you've done basically an audit on my website, digitalacomplice.com, and we're gonna show your process, really our process of like showing where the holes are, and then what are we gonna do about that to fill that with content? And let's just point out real quick, we're talking specifically about video content in my case, because I'm a you know video

producer by trade. This is what I do is to help people show up in AI search with video. that isn't always gonna be the way to go. So let's talk about why video can help and in what cases that is a good idea and in what cases that's maybe not going to be the right strategy. So we can kind of see if if this is like a good fit for people.

Adrian: Yeah. Okay. So digital accomplice as a case study, service business, B2B, selling the service of video production, et cetera, to businesses. So is it correct that your goal would be for when people are looking for video vendors, you want to get found for that.

Dane: Yeah, essentially. And I would even expand beyond that, sort of like a traditional video vendor is like executing on a plan that already exists. My my wheelhouse is really helping to develop those plans in the first place. In this case, specifically how to use video for AI visibility, as well as to expand video video strategic thinking to other kinds of video, because we're not just trying to show up in AI search. We're also trying to build relationships with trust. We're trying to trying to show products, we're trying to do

recruiting videos, event videos, there's all kinds of help that people need with video content that they're mo in most cases, most businesses are not well equipped to make really smart, efficient, sustainable video plans because they're not video experts. Lots of strategists and marketers, and there's lots of video production people over here, but I'm in that sweet spot right in the middle where I can help bridge those two gaps. And you know, there aren't a lot of people that do that. So it's probably not something that a lot of people are thinking about that they need.

Adrian: Beautiful.

Dane: But like they recognize some of the problems there, but they might not know that I'm quite the solution. So there's a little bit of category design or maybe like language problems. And that's why the GEO strategy is so compelling to me, is because now we have these sales transcripts. I've got buyer questions verbatim that I can though go plug those holes in my content plan, answering those specific questions that the buyers have. And I can do that in a way that most companies can't because I'm a video native and then now I'm working with you.

Adrian: Yep.

Dane: to get this really good data so we're gonna make data informed smart decisions and then measure the results and get the attribution to to drive further investment.

Adrian: What you're saying makes a of sense to me because everyone knows that video is powerful marketing medium and most of those people don't know exactly how to do it well and it's like it's a pretty intimidating unless you're like the kind of person who's like Ben who's like a quasi Instagram influencer where you're just like posting videos all the time and it's a very comfortable medium for you. I mean it makes sense to me.

Dane: Yeah, exactly. But that's only one kind of video, right? There's there's yeah, there's there's different kinds of video. There's high polish, there's low pol polish and authentic, and it's always changing. So like I spend all day, every day thinking and learning about this stuff. But like if you're a marketer or something like you're spending all your time on on other stuff too. So how are you gonna keep up with all these rapid changes and you know different use cases and the the evolutions of like social media went from being this like

Adrian: Yeah.

Dane: really flashy thing and people got kind of tired of that. Now they want authentic and you know, these things are always dynamic. So anyway.

Adrian: So everything you just explained, that whole answer you just gave provides the information we use to define your long tail and the written long tails that we're going to be optimizing for. Where once upon a time we might have been competing in Google for best B2B video agency. We're not, we're not trying to compete in like a high volume, poorly defined search category like that anymore. We're going to be digging into the long tail and all those.

use cases that you described and like having strategy in there. So one of our first steps is to

decide on a set of prompts that we want to be testing in the first place. And we actually start by doing traditional Google keyword research to figure out what there is demand for. then we take something like best B2B video agency and we build off of that. If we see that there's a lot of volume of searches for that thing in Google, okay, now we know there's demand for that topic. Now let's...

add the specificity to turn that into an AI prompt for us to be tracking. So here you can see.

Dane: Right. So the text is kind of small for for some people watching. So let's walk them through like what are we looking at here? This is the list of keywords and prompts in the long tail.

Adrian: Yes, so...

Dane: better, yeah. I mean they can always blow it up, but I just want to walk through just verbally what we're looking at here too.

Adrian: Okay.

Yeah. Okay. So this is a prompt. What are the best video strategists and video producers for BDB video and AI search visibility for BDB SaaS, AI, and cybersecurity companies in the San Francisco Bay area? Okay. Lots of specificity here. You have video strategy. That was important to you.

not just video production, video vendor, but video strategist. We've got B2B in there. We've got the AI search visibility goal of the video. We've got the ICP of B2B SaaS and AI and cyber security, and we've got the location of San Francisco. So we're kind of teaching to the test here because...

you're in San Francisco and it describes very precisely what you do. But that's actually what we want to do. We want to make sure for starters that you get found for a description of exactly the narrow thing that you do and who you are.

Dane: Mm-hmm. And where does this prompt come from at the at this time?

Adrian: we just came up with it and we think about, want to make sure we have coverage of the ICPs that you want to sell to and main solutions that you want to sell. And then if relevant, something like geography, you want to be in there one or multiple, and then whatever combinations of those that make sense to have.

So if you have three things that you sell and three ICPs, each of us could buy any of those three things. That's nine different combinations of different prompts we want to have to have coverage. And that's where that long tail.

Dane: Mm-hmm. So this is something we're engineering at the beginning of our testing based on what we do know about our ICP and the services and all that stuff. We're sort of like just putting a stake on the ground and saying, okay, this is what we're going after. So now we can start to get some baseline about whether we are or not showing up when people look for these things, right?

Adrian: Yep. Another thing we do when we're creating prompts is sometimes we will start by defining an ICP, an actual role or persona. The reason for this is I mentioned that an AI tool that you're using every day, like a chat GPT, has so much information about you. When people out there in the real world are asking questions, they're getting super personalized results.

Well, we want to try to simulate somewhat those personalized results that people are getting by defining the persona in the prompt itself. And that's what you can see here. I am CMO at a BDAI company. I am a VP of marketing at a B2B SaaS company. So that helps make sure we're testing for the right kind of people, testing your visibility for the right kind of people.

Dane: Mm-hmm.

Adrian: It's carving out little portions of like a broad category like this. One of the best B2B video marketing agencies.

Dane: Yeah, it's like a map, right? This is like just showing the kind of coverage we're looking to claim in our land grab.

Adrian: Yeah. Yeah.

So what are we learning? So visibility, the percentage of chats that mention Digital Acropolis, in this case in the last 14 days. So mention means the brand is mentioned in the generated answer.

To define terms, that's different from like a retrieval, which is of your actual URL or domain on your website or your YouTube channel or something. And that's when like a link is provided in the answer. That's a different metric. This is just if your brand is getting mentioned in the text. So this means that, so peak is running this exact prompt in...

Excuse me. Three different models every day. So I mentioned that Google is most important. We're testing both the AI overview and AI mode because they're slightly different along with ChatGPT. So Peak runs every one of these prompts through each of these three models every single day. And across all of those answers, Digital Accomplice.

comp list is mentioned 28 % of the time.

Dane: So for the naysayers about AI visibility measurement tools, like you know, there's n I I've heard some you know pushback and lack of consensus at times about like how scientific this is, because you can't really measure AI visibility because you get different answers. It's it's it's what probabilistic, not deterministic, right? So what what we're saying here is because of that, the methodology of measuring across these different platforms multiple times, like you know, every single day, that

Adrian: Yep.

Dane: that body of knowledge, bot body of data gives us some sort of like reasonable, defensible measurement system rather than just someone measuring on chat GPT on their personal computer. Is that is that the takeaway here? Okay.

Adrian: Exactly. Yep, exactly. And by testing different models, different prompts, running them every day, getting a large quantity of data, we can have some more faith that the kind of the average of that is 28 % across all of those is close to the reality that's being experienced out there in the world.

Dane: So then as like a gut check, twenty eight percent sounds good, but I see some red there and I just you know, not being the data science guy, based on what I'm trying to accomplish here, i is that a good thing what we're looking at there or is that concerning?

Adrian: That's not bad. Let's see.

The red has just changed over time, last 14 days versus previous 14 days. So invisibility did go down a little bit. The cool thing is that you're winning when you are mentioned, you're listed first. So average position 1.1 means you're almost always listed first when your brand gets mentioned.

Dane: Okay.

Dane: Okay, great.

Adrian: Yeah.

Dane: So so what is so what does this tell us what we need to do next?

Adrian: Yeah, so that's where it gets interesting. Let's dig into more detail.

Dane: Ha ha.

Adrian: We get so much data from this, it's kind of absurd. So now we're going look at a bunch of data just for this one prompt.

Adrian: First off, how do you rank compared to how visible are you compared to some other brands that we've chosen to track? These are selections we made that we wanted to compare you against. So for this particular prompt, you're the most visible.

It gets super interesting when we get to...

linked domains. So which domains are showing up in links in those generated answers. We'll see you right after.

Dane: Are we calling these citations or are these mentions?

Adrian: Yeah. a mention is just, is not a link, but demand that you are all mentioned as digital accomplice in text. Yeah. Yeah, exactly. And so one interesting thing we can learn here is that there's three domains here that have five links each while you have just two.

Dane: Okay.

Dane: Okay, so these are citations.

Adrian: which tells us that even though you're winning in visibility, you're not being referenced as a source as much by the AI. So that's probably where we're going produce more of the kind of content that's getting these three brands mentioned more.

Dane: So the takeaway from that is that they're producing more content, better content, more frequent content. What does that tell us?

Adrian: Let's see, it's probably more about the what kind of content.

So.

Adrian: URLs. I'm gonna look at some specific URLs.

Adrian: Let's go.

Dane: And this is a similar process you'd go through for any client, right? It's this is the the methodology.

Adrian: Yeah, totally. Yep. So I've zoomed out now, we're looking at all prompts, not just that one prompt and moved over to a different part of the tool, URL specific citation data. And so we can see right here which exact URLs are being retrieved most often across all those chats, across all the prompts.

All the chats every day, three different models. A retrieval is when that URL is, a answer includes that URL, a citation is how many times. So sometimes one AI answer might actually cite the single URL multiple times across its answer. And that would generate one retrieval with multiple citations.

Dane: Mm-hmm.

Dane: Okay.

Adrian: So if you want to really see the impact that a piece of content is having across the answers, you could look at total citations.

Dane: Let's let's talk for a second about the the value of a citation. You know, I think I'm still trying to wrap my head around like what's a good thing to happen. It's like I think we want to show up in AI search, but there's different ways to do that. You could be mentioned, you could be cited. I think ultimately we w wanna be recommended is the way to drive revenue. But you said something to me once is like we we take it as an inherently good thing just showing up in the first place. And to me, this sounds an awful lot like we want to be part of the conversation.

Of the decision of whether or not that we're a solution to our clients' problems, right? And so being being listed right there, I mean, if you just got this little link right there, but the AI overview is scraping your site and giving them the answer and then they get the answer they need and they leave and they don't click on that link. Is it still valuable? I mean, probably, but if they're not clicking it, how valuable, right?

Adrian: Yeah. Yeah. So there's a little bit of value there. It's a, it's a free impression. and let's not discount impressions, you know, in this kind of performance marketing era where we want to support everything with data and actually read everything, you know, when Nike pays for a Superbowl ad, they attributing precisely sales to that ad?

Not exactly. They know that eyeballs on their brand helps their market share. So impressions are inherently valuable. But yeah, I mean, you make a video, you'd like people to actually watch the video, not only have it influence an ad answer or give your brand a glimpse of exposure and an answer.

And so that's where I think it comes down to the kind of questions that people are asking. Is it the kind of question where AI is really answering the question, it cites whatever sources, answers the question, the engagement is done? Or is it the kind of question where the person's probably going to actually pick one of those video citations to go watch?

Dane: And we're I I think my sense is that those two different outcomes are related to the buyer's sales funnel, right? It's like the earlier things, the questions that are like, what is this thing? Those are more likely to be answered by an AI overview without someone going deeper. Whereas when people get later in the stage and they get more commercial intent, they really want to dig into like what the what the deal is so they can make a buying decision. Does that track with what the expectation is? Okay.

Adrian: Yeah.

Adrian: Yeah. I think a lot of times the behavior that we see is AI shortens the list of options that someone considers and then they go dig deeper into one, two, three alternatives. And that's where actual click throughs happen. That's where they jump over to Google and do a traditional search and go find a website and they're watching videos and everything about those.

one, two or three leading candidates.

Dane: And my sense has been that you the benefits of video content showing you know putting video content out there versus text content is that the text content is more likely to get scraped and summarized without citing you or giving you any credit and you remain invisible, whereas the video is more likely to show up like in those top results with like the thumbnail, and you know, you're getting a little bit more visibility, sort of like forcing yourself into those results as opposed to.

letting them eat your lunch and you don't get any credit or mention. Is that does that track with like how we should look at this?

Adrian: think that's roughly... probably right? I don't know.

Dane: Yeah, we have to be cautious about this being always true. Like we've we've seen that video is not always the answer, but in s in many cases it can be. So we should get into that here too.

Adrian: Yeah, I mean, I mean, yeah, let's, let's actually dig into that question. So you've learned about your visibility by putting in a bunch of prompts and you see a result like, like this brands ranked by visibility, digital accomplice sitting at 2%. it was 28 % for one prompt and like 3 % for another, and then not mentioned at all across any of other prompts we're testing.

Okay, we want to increase that. What do we do? Well, let's go figure out what type of content is influencing AI answers in this space. Now you will have heard headlines about what websites, what domains get cited by AI the most overall and that's it varies by the different models. But overall, it's like Reddit, YouTube, LinkedIn, Instagram.

Forbes is up there. I think that the National Institutes of Health government website is high on the list for one of the models. There's some variance medium, I think it's pretty high on the list. That's overall across all different use cases and the myriad ways that people use AI. It's much more helpful for us to look at what type of content is influencing AI answers in your space.

for the types of prompts that you believe your audience is actually putting into AI. That's what we're testing here. And that's where we jump over to look at the domains being retrieved the most. So now again, we're looking across all the prompts that we're testing and we want to see at the domain level what's being retrieved, which is a way of seeing what is influencing AI's answers and what is...

gaining the opportunity to get some click-throughs through the AI. So this phase is YouTube. For the prompts that we're tracking, let's actually, the audience's sake, just scroll through the entire list really quick so you can see what we're tracking.

Adrian: VP Marketing, B2B SaaS, Video Production Companies, VP Marketing, 50 person, B2B SaaS, Done For You, YouTube Help, AI Search, GEO, Cyber Security, Marketing Director, Professional Services Firm, How Do I Get My Company Set In Chat, GBT, Using Video, Product Marketing, Explainer Video, that's a more narrow type of video, Marketing Director,

Founder of SAS.

So lots of different variations. So without a mind, that's what we're testing. Across that set of prompts, YouTube is the domain that is retrieved the most often.

then it's a lot of these competitors.

Dane: Interesting that LinkedIn is down LinkedIn is down at the bottom. And you know, I'm I'm I'm most active on LinkedIn and YouTube. So that's interesting to me that they're so spread apart.

Adrian: Yeah, but also like LinkedIn is really a factor. You know, it's almost half as many retrievals as you two. It's definitely a factor. And then this looks like a lot of your competitors are rough competitors. Retainable.com, I don't know what that is. Let's find out really fast. Retainable.com. Give your clients a better retainer experience and scale your agency.

okay. They're like consultants for agencies, apparently. Well, our goal here is to figure out what type of content influences the AI so that we can decide what type of content we want to produce to increase our visibility. Now we know what domains are the most influential. Let's go look at the actual type of content. Drill down into one of the domains.

we can go see the precise URLs that are getting retrieved. And so what I would want to do for you is look at like these top five, especially that top one that's an outlier.

and take a look at it. What is the topic? What is the length? Now it's the time for you as a video expert to kind of assess this content and see what you think of it, what you can learn from it.

and do that for the top. Go ahead.

Dane: And what what do we Yeah, just high high level, what do we think like a a strategy would be? Like if we see that the content that is showing up, am I trying to unseat them and you know be the one that has better content on that topic? Or am I trying to go after the holes that aren't filled yet? What's the what's the playbook look like? And

Adrian: We're trying to follow trends in what works and what AI is actually referencing. if there... Yeah, there's probably some combination of competing directly with content that's already out there and finding gaps. Basically, we've affirmed that YouTube is influential in this space.

So it's worth working on video content to achieve your goals. Having established that, I think we picked the types of videos, the topics, based on which of the prompts are highest priority and least visible for you right now. And produce that video, whether or not there is a good video on that precise topic already out there.

Dane: Yeah, how how do we know

Dane: How would we know which of those prompts would be priority for me? Is that based on the services I'm most interested in selling or the things that most people are wanting to buy? Or how does that work?

Adrian: Yes.

A little bit of both. Definitely those are the two places to start. What do you most want to sell and what is there the greatest? Google search volume. Going back to how we started this research by looking at Google search volumes. You know, we want to test. We don't put things into Google as long as this, but we could do like B2B video, San Francisco. could do some slightly less narrow searches, but that are a long tail in Google terms.

Dane: Mm-hmm.

Adrian: NC volumes there that kind of measures demand. We also are getting estimates of AI volumes from these tools like peak. So let me just turn this on.

It's just simply a five point scale. It's relative.

Dane: Mm-hmm.

Adrian: One is very low, five is very high. We can sort on that.

Dane: Okay, so my my hypothesis is that what I really think people need to buy and what I really want to sell is a monthly retainer for creating video content for these companies that are answering their buyers' questions and then sell you know doing a a process just like this where we're measuring where those holes are, we're producing the content, we're showing them how it gets plugged and how you can attribute that to the revenue. And we can also like get attribution from the sales people.

Like, are people showing up more qualified because they're consuming this content and they're getting their questions answered? So they're not asking those questions anymore. Cause ideally, those people that we've educated are self-educating, and then they show up when they're ready to take a software demo or talk to a salesperson, they're like, they've already kind of made their decision that this is the right thing for them. So that's really what I'm hoping to sell, because I think that's what's most important and helpful for companies, moving from an SDO only era to a GEO era.

If we're not creating that video content that answers buyer questions, we're kind of like le leaving money on the table. We're leaving that open for the competition to eat our lunch. So I think that's the most valuable thing. I'm just still not clear on what magic words I need to be using to to make sure I'm plugging the right hole to get found for that stuff. What is what does this tell us about how I would fulfill a strategy like that?

Adrian: I mean so we could use peaks estimate of volume here and prioritize these ones that are like getting threes and fours So like this one is a pretty good title like an imagine for YouTube video How do I get my company cited and chat you between who have is using video? Maybe picking just one of those AI models in the title

Dane: Mm-hmm.

Adrian: Cybersecurity is very narrow, it's three out of five moderate search volume for this keyword. So you love to see like a long tail that still has some volume.

Dane: Yeah.

Dane: But it's not quite clear then l exactly which video I should make next. We're still sort of like guessing a little bit or hypothesizing.

Adrian: Adrian: That's where your taste comes in. You're the video guy.

Dane: Yeah. Right.

Adrian: That's where content marketers come in. You know, my role, I do a lot of website content. We figure out topics and then we apply some taste to decide how to address that topic, what the piece is actually going to look like. But part of it is digging into the URLs that are being cited for this prompt today and looking at patterns.

Dane: Mm-hmm.

Dane: Mm-hmm.

Adrian: across those, including the non-video content.

Dane: Okay. So are we at a place where we can say what a good video strategy would look like for me and my next batch of videos that I put out there?

Adrian: Okay, yeah, let's decide the next video for Digital Accomplice to produce.

Let's start with the problems that we want to optimize for. Is any of you standing out to you? How about cyber?

Dane: Cyber's good, yeah. I've done some cybersecurity video content, but as a you know as an industry, they're very risk adverse. And doing video content is I think concerning for them because they don't want to have like a deep fake made of them and have their likeness out there. So it's an it's an interesting con conundrum. But I think getting cited in ChatGPT seems like that's I think every CEO is probably thinking that or asking that.

Adrian: Yeah, right.

Yeah.

Dane: Why are my competitors showing up? How do we get cited?

Adrian: Okay. Yeah.

They're not all including the using video part in the question, but some will.

Dane: Exactly. They're problem aware, not solution aware for video.

Adrian: Yeah.

Adrian: Let's go with this, the highest volume prompt that we're Marketing director, professional services firm, looking for video production companies.

Dane: Mm-hmm.

Adrian: Alright, clicking into that.

Adrian: Today you're not being mentioned.

Adrian: Let's look at links.

Adrian: Actually, let's look at an example first.

Adrian: You can actually see the detailed data that we can drill down into. So here's the actual interaction with the AIO review. There's the prompt. Here's the answer it got back.

Dane: So I'm thinking the way to do better than that is just somehow go deeper and make a better answer than this answer. More data, more depth, more context, something like that.

Adrian: Well, this is AI's answer, so what we want to look for is what...

sources, it's exciting. So we've got a blog about best video production companies, another ranking of best B2B video production companies, so website content is actually what's for this prompt, which is often the case for best services, best firms type of search, best providers will be...

Dane: Mm-hmm.

Dane: These are frequently listicles, right?

Adrian: Exactly, that's a listicle. this one's actually not as much video content that's influencing and showing up in the answer. But you have a website, so you can go make a listicle that's maybe a little bit more long tail than these. So you could say like best video production companies for the best strategy video or best video strategists or something like that. Best B2B video strategists.

Dane: And then in that listicle, I'd be listing the competition as well, but I'd be putting myself favorably. Yeah.

Adrian: You

Adrian: you would, but you kind of like pick the criteria for analysis that helps you look strong.

Dane: Yeah, and to me that's a differentiated value question is like, you know, I'm not the right fit for everybody, but if you are in the San Francisco Bay Area, so am I. So that's probably I'm probably a better fit than someone who isn't.

Adrian: Yeah. And I'll help you with your actual studio because I can come on by. Yeah. Let's see. guess there is some YouTube that comes up sometimes, so let's take a look at that.

Dane: Yeah, exactly.

Adrian: top 10 creative agencies of video production.

So it's coming from a video production company's channel that has 16,000 subscribers. This video has 500 views. It was posted six months ago. Let's watch a few seconds.

Dane: Yeah.

Adrian: Maybe not.

Dane: Come on, internet.

Dane: But we're looking to see why is it performing, see if we can glean any intel.

Adrian: Okay, I do not just do so. But now we're getting to...

your expertise where you start to look at this and see, could I make something like this? Could I make my own version of this that will reach my people specifically?

I'm sure you could create.

Dane: Yeah, being be how can I be more helpful to my buyers, basically?

Adrian: Yeah. Yeah.

Dane: So then I would come up with a plan, make that video, put it out there, and we'd check back in after some amount of time to see if we've seen movement.

Adrian: Yes, but I want to do a better job answering this question of what your next video should be. So let's look at the most retrieved YouTube.

videos across all the prompts we're checking.

So now this is across all prompts, just YouTube URLs. Let's go one at a time.

Adrian: this is the same one.

Adrian: So that is actually the most, sorry, I didn't share my other tab. It is that same video that is the most cited video across all of the prompts we're tracking. So it turns out that a listicle in video form helps influence answers about a vendor.

Dane: Okay. So that sounds like that might be my next video is me comparing me to all the other options that I can think of that might be relevant to those prompts and and really help the help the buyer evaluate whether I'm the right fit and do so in an honest way and in depth enough that they're really equ I'm equipping them to make a a a good decision rather than just pumping myself up and saying how great I am.

Adrian: So.

Adrian: Yeah.

Adrian: Yep. Yep. And by the way, that's really effective for AI influence. When you're super clear about what you do, what you don't do, who you serve, or you don't serve. When you define your niche well and you're consistent in how you describe your niche, AI knows when to serve you and when not to.

Dane: Mm-hmm.

Dane: So I think for a lot of business people, this is going to be a departure from the way we've done it for so long, is like you know, saying good things about yourself is kind of like that's business as usual in business. But what now we're saying that if you want to show up, you you can't just be marketing fluff. You've gotta help them kind of break apart when you're the things that you might not be a good fit for as well.

Adrian: you

Dane: So you can't be everything to everybody, is what I'm hearing.

Adrian: Amen. Totally true. Let's look at the number two video.

Dane: Yeah.

Adrian: you

Adrian: Hmm.

Adrian: Okay, so this is the number two sided YouTube video across all your prompts. It's got 3,000 views, channel's got 22,000 subscribers.

Adrian: Yeah, another good source of inspiration for you.

Dane: Okay.

Adrian: And it's the same type of thing. it's, I mean, it looks like he's just talking about himself, his own agency, but he's specifying explainer video for BDB SaaS. It's like quite narrow in two dimensions. The solution and the ICP are both defined here.

Dane: Mm-hmm.

Dane: Okay. So I I guess I'm starting to think that we're getting to a conclusion here about what the what the process you have working with Peak and doing this research, how that's going to tell us what the game plan should be and like what we're going to do about this. So we're kind of like if another company is watching this and they're thinking, okay, if I go through this process, I will then see what's showing up and what's going to be the best, you know, alternative best plan for me to go.

Adrian: you

Adrian: you

Dane: and and create the content that's gonna give me a shot of showing up and then measure it and see if it's happening.

Adrian: Yep. And then track over time.

Dane: Okay. what else do they need to know about this part of the process? Like, I don't know, just kind of looking for a wrap-up idea here is like we've kind of walked through how to find out where you stand. And now we've talked about what to do about it, and we've talked about how to measure that. I guess the the big takeaway I'm looking for here is like how can we generalize what this process

Will do for us to use data to inform further investment, right? How what how should I know how much money to put towards this to accomplish my business objectives? And maybe for me, we could say, hey, I want to close one new deal. What's the so what's what's the amount of effort I should put in here to close one new deal? Is it is it that cut and dried, or we don't know?

Adrian: I don't think I can answer that for you.

Dane: Okay. All right. Well

Adrian: Yeah, do you want to ask that a different way?

Dane: Yeah, let's let's say can we summarize what we just did and what is the story that we're telling here about why this process is important to showing up in AI visibility?

Adrian: Adrian: Sorry, I'm drawing a bit of a blank. think my, 12.08 AM Bali Indonesia time is catching up with me.

Dane: That's okay.

Dane: Yeah. Right. Well, all right. Well, here's here's what I think we learned. it's now possible to measure pretty reliably and reasonably where you show up in AI search and where you don't, and to use that data to make a content plan that will then measurably show up, show you how you the the improvements that you've made and whether you are actually moving the needle to show up.

Adrian: Yep.

Dane: Yeah. All right. Well, if people want to learn more, you know, reach out to me, Dane Frederiksen at DigitalAccomplice or Adrian Dahlin at search to sale and we'll walk you through the process. Bless you.

Adrian: Excuse me, sorry. Sorry I'm doing a terrible job of helping you land this plane. Well, we covered a lot. Yeah, I'm calling in from Bali right now. We're on traveling. So it's very late at night.

Dane: No, it's okay. You know what we're doing is we're being authentic. We're building trust with authentic authenticity. Which is

Dane: Yeah, exactly. Yep. Well, thanks Adrian for walking us through that. I'm sure we'll we'll be checking in to to find out where we stand and continue optimizing.

Adrian: Are you going to go make a video about the best video agencies for video consultants or video strategists or GEO?

Dane: It it sounds it sounds like I should, yeah.

Yeah, it sounds like I should. I guess, you know, the question on my mind has been like, who are my competitors? Am I competing against every video production company or other video agencies or only people doing GEO content? And I just happen to be doing it with video and they're doing it with text. So I mean, you could say all of those are my competition.

Adrian: I mean, would dig into this data to help answer that, see what other agencies are showing up.

Dane: Yeah.