Episode Summary
Araminta Robertson, founder of Mint Studios, walks through the GPT article framework, a methodology for improving brand visibility in AI tools like ChatGPT and Perplexity.
Key Takeaways
- 1LLM visibility tracks how often your brand appears in AI tools like ChatGPT, Gemini, Perplexity and Claude, and now drives 30-40% of inbound leads for some B2B brands.
- 2The GPT article framework is a tested methodology: select bottom-of-funnel topics, create 5-10 prompts per topic based on customer understanding, then write one 500-1,000 word article per prompt and publish on your own website.
- 3Case study results show visibility rising from 5% to 34% in under a month, with smaller brands cited alongside competitors like Stripe despite vastly smaller budgets.
- 4Self-attribution remains the most reliable tracking method because users learn brand names via ChatGPT but type them into Google rather than clicking citations, so LLM-driven leads often appear as direct or organic in analytics.
- 5GPT articles complement rather than replace SEO, since around 70% of users check both Google and LLMs during research, and many GPT articles rank organically as well.
Contents
- Introduction
- What Is LLM Visibility and Why Does It Matter?
- How to Improve Brand Visibility in AI Search: The GPT Article Framework
- Which LLMs Should You Target for Maximum Visibility?
- How Do You Pick the Right Prompts to Target?
- What Makes a Good GPT Article? Best Practices for 500-1,000 Word Content
- Where Should You Publish GPT Articles?
- Case Study: How Fiska Increased LLM Visibility from 5% to 34% in Under a Month
- How to Track LLM Visibility and Measure Success
- Do You Still Need SEO If You’re Doing GPT Articles?
- How Future-Proof Is the GPT Article Framework?
- What This Means for Your Marketing
Introduction
Large language model (LLM) visibility is no longer experimental.
For B2B brands, particularly in fintech and financial services, between 30 and 40% of inbound leads now arrive via large language models like ChatGPT, Gemini, Perplexity and Claude.
This is not a niche channel. It is a major acquisition source, and one that most brands are not yet deliberately targeting.
How the Framework Was Built
Araminta Robertson, founder of Mint Studios, developed the GPT article framework in mid-2025 to address exactly this gap. Her agency tested digital PR, LinkedIn articles, and multiple content approaches before landing on the methodology that consistently delivered results: short, product-focused articles targeting specific prompts.
The case study numbers are striking. Fiska, an embedded payments company, increased visibility from around 5% to 34% in less than a month. The brand began appearing alongside competitors like Stripe, despite operating at a fraction of the budget.
What Is LLM Visibility and Why Does It Matter?
LLM visibility tracks how often the name of your brand appears in responses from large language models. Because these tools generate answers probabilistically, visibility is measured as a percentage: in what share of relevant prompts does the LLM name your brand?
The shift matters because of scale. Around 70% of users now check both Google and LLMs during research, and for some brands the proportion of leads originating via AI tools has reached 30 to 40%.
The scale of LLMs as a customer acquisition channel
“I think it’s huge. We have clients where 30-40% of their inbound leads come via LLMs. So this is not just one little channel. This is clearly like a big acquisition channel.”— ARAMINTA ROBERTSON, FOUNDER, MINT STUDIOS
Attribution, however, remains messy. Very few users click citations inside ChatGPT or Gemini. Instead, they learn a brand name via the AI tool, then type it directly into Google.
This means LLM-driven leads often appear in analytics as direct or organic traffic, making traditional channel reporting unreliable for measuring AI impact.
How to Improve Brand Visibility in AI Search: The GPT Article Framework
The GPT article framework is a repeatable, testable methodology for increasing brand visibility across LLMs. It emerged from extensive testing of alternative approaches, including digital PR, LinkedIn publishing and structural content changes, none of which delivered comparable results.
The model works in five steps, each addressing a specific part of the visibility challenge.
Step 1: Select Bottom-of-Funnel Topics
Begin with topics where the researcher is already actively looking for a solution. For an accounting tool, this might be “top accounting tools for designers” or “QuickBooks vs Xero comparison”.
These are not awareness-stage topics. The user has already identified a problem and is evaluating options.
Step 2: Create 5 to 10 Prompts Per Topic
For each topic, construct a set of prompts you believe users are typing into ChatGPT, Gemini or Perplexity.
Why educated guesses about user prompts work better than exact semantics
“Once you have your topics, we then put together 5 to 10 prompts that we think that user is typing into ChatGPT, Gemini. So yes, these are mostly guesses. But they are usually based on a deep understanding of the ideal customer.”— ARAMINTA ROBERTSON, FOUNDER, MINT STUDIOS
The exact wording matters less than targeting the right general phrase. Appearing for one prompt tends to increase visibility for similar variations.
Step 3: Write One Article Per Prompt
Each prompt receives its own article, typically 500 to 1,000 words. These pieces are shorter than traditional SEO content, reflecting the narrower focus of each prompt.
The tone should be factual, to the point, and product-focused. Go into detail about the product itself rather than broader industry commentary.
Step 4: Publish on Your Own Website
Testing showed that publishing on the client’s own website outperformed LinkedIn, Medium and other third-party platforms. It also avoids the awkward optics of writing bottom-of-funnel comparison content on a platform designed for thought leadership.
Step 5: Track Visibility and Measure Leads
Use LLM tracking tools to monitor what percentage of your target prompts now mention your brand. Then measure actual inbound impact via self-attribution, since analytics alone will not reliably surface AI-driven leads.
Which LLMs Should You Target for Maximum Visibility?
The priority LLMs are ChatGPT, Gemini, Perplexity and Claude. Google AI Overview, a variation of Gemini, also matters for brands with strong organic search presence.
Claude is more expensive to track due to its technical setup, but the other three are accessible and widely used across B2B research workflows.
How Do You Pick the Right Prompts to Target?
LLM tracking tools like Peec AI and Profound provide volume indicators, but these should be treated as rough guides rather than definitive data.
The more reliable method is to understand your ideal customer deeply: their pain points, their problems, the questions they ask when evaluating solutions. Then use common sense to validate whether a given prompt is genuinely plausible.
You can also use ChatGPT or Claude to brainstorm a long list of potential prompts, then select the ones that make the most sense for your audience. The AI generates options; you make the judgment call.
Why Exact Prompt Matching Doesn’t Matter
“It doesn’t really matter the semantics. Obviously there’s millions of variations of how someone might do research, but what we found is that if you generally target that phrase, you will appear for similar phrases as well.”— ARAMINTA ROBERTSON, FOUNDER, MINT STUDIOS
What Makes a Good GPT Article? Best Practices for 500-1,000 Word Content
A GPT article is not a long-form SEO pillar post. It answers a specific, narrow question in the most direct way possible.
Length sits between 500 and 1,000 words because the question itself is narrower. You do not need the same depth as a 3,000-word guide.
Formatting Elements That Improve LLM Parsing
| Element | Recommendation |
|---|---|
| Headers | Descriptive, clear, summarising the content beneath |
| Bullet points | Use liberally to break up dense information |
| Tables | Particularly effective for comparisons and data |
| Synonyms | Include natural variations of key terms |
| Writing style | Factual, to the point, not storytelling |
These best practices appear to have a marginal impact. The most important factor is simply having a piece of content that answers the question clearly and directly.
Where Should You Publish GPT Articles?
Publish GPT articles on your own website, not on LinkedIn, Medium or other third-party platforms.
Testing showed that LinkedIn and Medium underperformed compared to owned domains. LinkedIn also presents an optics problem: bottom-of-funnel comparison content reads strangely on a platform designed for thought leadership and network engagement.
Your own website gives you full control over structure, speed and indexing, and avoids diluting authority across multiple domains.
Case Study: How Fiska Increased LLM Visibility from 5% to 34% in Under a Month
Fiska, an embedded payments company, began with overall LLM visibility of around 5% for relevant, high-intent topics.
After publishing GPT articles following the framework above, visibility climbed to 34% in less than a month. The brand began appearing in responses alongside Stripe and other competitors with vastly larger budgets.
Crucially, visibility was not the only metric that moved. Fiska also saw a larger influx of inbound leads, confirming that the visibility increase translated into actual business impact.
How to Track LLM Visibility and Measure Success
LLM tracking tools measure what percentage of target prompts mention your brand. Tools like Peec AI offer clean dashboards and recurring competitive reporting. The founders innovate constantly, and the UX remains simple even as features are added.
However, tracking visibility is only half the picture. Measuring actual lead impact requires a different approach.
Why Self-Attribution Is the Most Reliable Method
Very few users click citations or source links inside ChatGPT, Gemini or Perplexity. Instead, they learn a brand name, perhaps do a little more research, then type the brand directly into Google. HubSpot and similar platforms label them as “everything except AI referrals”.
The most reliable tracking method is simply asking users directly: “How did you hear about us?” Self-reported attribution cuts through the messy data and surfaces the actual source.
Do You Still Need SEO If You’re Doing GPT Articles?
Yes. GPT articles and SEO work together, not as alternatives.
Around 70% of users check both Google and LLMs during research. Attribution is messy precisely because buyer journeys now span multiple channels, and it is difficult to know definitively whether someone found you via an LLM, via Google, or via both.
Many GPT articles also rank organically, delivering value through traditional search as well. The writing is high quality, the topics are specific, and the structure is clear. These are exactly the qualities that perform well in Google as well as in AI responses.
SEO work gets you halfway there. Adding GPT-specific optimisation and more bottom-of-funnel content makes a substantial additional difference.
How Future-Proof Is the GPT Article Framework?
Predicting the future of LLM behaviour is difficult. The models could change their sourcing methods, their ranking signals, or their entire approach to generating answers.
However, the framework has a built-in durability. LLMs still rely heavily on scraping Google’s index, and Google has decades of experience indexing information. There is no comparable alternative.
“I mean, that’s hard to say, right? I cannot predict the future. All I can say is that we have to adapt.”
Likely Scenarios and Content Durability
The only scenario that would fundamentally disrupt this approach is if LLMs stopped scraping Google entirely and moved to a fully closed model. That would not produce a good product, so it seems unlikely.
In the meantime, the content you create is high quality, specific and useful. Even if LLM behaviour shifts, the articles still perform organically and still convert users who find them via traditional search.
What This Means for Your Marketing
If your customers use AI tools during research, LLM visibility is now a genuine acquisition channel. It is not speculative. It is measurable, testable and repeatable.
The GPT article framework offers a practical starting point: pick bottom-of-funnel topics, create 5 to 10 prompts per topic, write one short article per prompt, and publish on your own website. Track visibility with tools like Peec AI. Measure lead impact via self-attribution.
This complements your SEO work rather than replacing it. Both channels matter, and the buyer journey now spans both.
The brands that move early will build visibility while the channel is still underutilised, and the methodology works whether your budget is large or modest.
Read the Full Episode Transcript
[00:00] Araminta Robertson: LLM visibility tracks how often the name of your brand is appearing in LLMs. I think it’s huge. We have clients where 30-40% of their inbound leads come via LLMs. So this is not just one little channel. This is clearly like a big acquisition channel.
David Bain: How can companies grow LLM visibility and acquire more customers via the GPT article framework? Let’s find out with Araminta Robertson from Mint Studios.
[00:20] David Bain: The Pump and Funnel Marketing Show with David Bain. Hey, it’s David. Way back in 2007, I delivered my first ever digital marketing training session. Ever since then, I’ve been on a journey to help professional marketers pull together every disparate marketing activity into one unified model, the Pump and Funnel marketing model.
[00:44] Araminta, where does LLM visibility fit into the Pump and Funnel marketing model?
Araminta Robertson: Hi David, good to be here. I would say it fits in the Help or the Heart section. So GPT articles are long-tail short articles that answer very specific questions as part of a topic.
[01:02] And these are questions that go into more detail about a specific topic. For example, if we’re talking about accounting tools, it would be, you know, compare these two accounting tools, or how does this accounting tool help me do XYZ? So I would say that, yes, therefore helping the user along their journey. And also maybe the Heart, because there’s a lot of product information usually in these articles.
[01:30] David Bain: That’s what GPT articles are. And in terms of LLM visibility, what is that exactly? Is that simply some element of your content appearing within AI search?
Araminta Robertson: That’s right. So LLM visibility tracks how often your brand, the name of your brand, is appearing in LLMs like Gemini, ChatGPT, Claude, Perplexity, all of these LLMs.
[01:57] So it’s tracking in a response how often, because it’s a probabilistic kind of type of response, 80%, 50%, is the actual brand name appearing in the response of the LLM.
David Bain: Got you. Okay, and you mentioned a few LLMs there. Are there any particular LLMs that are absolutely key at the moment to appear in?
[02:23] Araminta Robertson: Yeah, I would say ChatGPT, Gemini, Perplexity, Claude as well, but it’s a lot more expensive to track Claude because of their setup. But those are the main ones, I would say. And Google AI Overview, which is kind of Gemini but in a different model.
David Bain: Okay, so I’m sure we’ll cover how to track later on, maybe in the software that you’d recommend.
[02:41] But in terms of the GPT article framework that you mentioned, do you want to give a little bit more meat on that in terms of what it is and perhaps how to pick the prompts to target initially?
Araminta Robertson: Yeah, so the GPT article framework is a model that we came up with middle of last year, 2025, when, you know, more and more of our clients were trying to figure out like how do we improve our brand visibility on LLMs.
[03:08] So we tested a lot of different methods. I have an article that goes in detail about all the different approaches we tested. So we tested, you know, digital PR, we tested LinkedIn articles, different— changing the way we write our articles, all these things.
[03:27] And after doing a lot of testing, what we found worked better was actually creating more content, but that is very specific, specific to the phrases, the prompts that you’re targeting. So the model starts off with deciding what are the topics that you want to target, and you start with ideally very bottom of the funnel topics, topics where the person who’s researching this is already actively looking for a solution to their problems.
[03:49] So if we go back to the accounting tool example, this person would be looking up, you know, top accounting tools for designers or QuickBooks versus Xero, those kind of very bottom of the funnel topics. Once you have your topics, we then put together 5 to 10 prompts that we think that user is typing into ChatGPT, Gemini.
[04:10] So yes, these are mostly guesses. But they are usually based on a deeper, a deep understanding of the ideal customer. And it doesn’t really matter the semantics. Obviously there’s millions of variations of how someone might do research, but what we found is that if you generally target that phrase, you will appear for similar phrases as well.
[04:30] So once you have your prompts that you’ve put together, you then create one article per prompt. And these are short articles. Usually 500 to 1,000 words. And there’s a few best practices when you create them. They want to be quite factual, quite to the point, descriptive, ideally product-focused. You go into detail about the product and a few other best practices to ensure that LLMs basically scrape them.
[05:00] And so with the GPT article framework, we have found when tested across all the other methodologies, this is what time and time again increased the visibility the best. For our clients. And I can share a few examples of how we help some of our clients go from like 5% visibility for the prompts that we were targeting to over 40%.
[05:18] And not just did the visibility grow, but also we saw an influx in inbound leads. So there was an actual impact on business results. So that’s a bit of a summary.
David Bain: That’s absolutely amazing. I’d love to hear some examples as well. I’ve got 3 notes of different things that you shared already, actually.
[05:36] So I just got to clarify a couple of things. And to begin with, you talked about LinkedIn articles, testing LinkedIn articles and others in terms of best places to publish. And then you said, you know, it all comes down to the actual prompt itself. Are you saying that it doesn’t actually necessarily matter where you publish the piece? It’s all down to how you go about phrasing the prompt that you’re attempting to target?
[06:00] Araminta Robertson: No, I would say it does matter. We publish the articles on the client’s website. So on your website, that is the best place to publish them. So that’s what I was saying. We tested LinkedIn, we tested Medium, and first of all, it’s really weird to write a bottom-of-the-funnel article piece on LinkedIn. So just also doesn’t— the optics aren’t great.
[06:18] And then it just didn’t have an impact as much as— even though everyone says LinkedIn keeps appearing as one of the biggest cited sources, it’s true. But in terms of when you’re influencing the LLMs, we found that just publishing content on your own website is what works best.
[06:30] David Bain: And you also talked about guessing the prompts, obviously intelligent guesses based upon knowing customers intimately. But is there any software out there? Are there any resources that you can actually use to determine what are likely to be more successful prompts?
Araminta Robertson: I mean, these LLM tracking tools like Peek and Profound, they will have a little dot, like a little kind of chart that will say the volume.
[06:55] But I wouldn’t take those too seriously. I think it comes down to really understanding your ideal customer, their pain points, what are their problems. And it should kind of, using common sense, like, yeah, it makes sense that they would search this.
[07:12] And then don’t worry too much about the semantics, you know, in the sense of like, how do they write that question? What matters is, are you targeting the topic? So I would— you could use ChatGPT or an LLM, and that will help you, like, for example, brainstorm with Claude or whoever, what are the questions that they might search.
[07:26] And then with the LLM, with Claude for example, you can put together 20 prompts and then you’re the one who picks out like, I think this makes sense, I think that this makes sense. So you can use an AI to help you put together the prompts, but ultimately I think the person should be the one to decide what to target.
[07:46] David Bain: And you also mentioned 500 to 1,000 words, why is that the sweet spot? And would you like to maybe clarify a few different elements that are ideal to include within the article as well?
Araminta Robertson: Yeah, I would say it’s a range. So it doesn’t have to be 500 or it’s just, I would say just shorter than usual.
[08:05] Because I mean, I guess for us, the longer ones are more for SEO. I think with shorter ones, you just don’t need to go as in depth, because these are shorter questions that you’re answering. And some key things that I would include, so I would make sure that your headers are quite descriptive.
[08:17] So make sure that there’s a whole list of bullets. We have an LLM optimiser that we use where we’ll create the piece and then we’ll put it through that. And there’s a few things that it’s gonna try and do. So put lots of synonyms, make the headers descriptive. It really likes bullet points, tables. Tables are a big one.
[08:40] Make sure that the writing is, yeah, to the point and isn’t like super storytelling. What else? Yeah, I would say those are the main ones. There’s a bunch, like a full list of bullet points and it’s in the article that I put together explaining our methodology. And these are, you know, it’s funny because at the beginning we thought they had a big impact, but we’re finding they have less and less of an impact.
[09:03] So I think the most important is that you have just a piece of content that answers the question. That’s the number one thing. And then all these little best practices, I think they might have a small impact, but they’re not like the most important thing.
David Bain: Lovely. Okay. And you mentioned case studies as well. You know, what brands have been successful at doing this? And how do you actually track success?
[09:26] Araminta Robertson: Yes, so I mean, for external brands, I’m not sure because each one has their own methodology. But for our own clients, we’ve had quite a few kind of success, success stories that we can share. So for example, I’ve got the notes here, so I can name one of our case studies, Fiska.
[09:39] They’re an embedded payments company. When we started, their overall visibility was like about 5%. For really relevant and good topics. After publishing GPT articles, within less than a month, their visibility started to grow, and to like 34%, which was huge. And they were outcompeting other companies like Stripe and others who have huge budgets.
[10:03] And not only was the brand Fiska being mentioned a lot more, but it was cited a lot more as well, next to companies like Stripe and all that. And so now if you look up embedded payments-related topics on Gemini, ChatGPT, Perplexity, Claude, you’ll find that Fiska turns up a lot more.
[10:16] And that was following our approach of picking topics, creating GPT articles, just that. And that worked really well. And not only did that happen, but they also started to see an influx or a bigger influx of inbound leads. Interestingly, it’s hard to track these things. And the best way we’ve found is by asking the user, how did you hear about us? So self-attribution.
[10:41] Because they’re often marked as direct. Organic, like literally everything except AI referrals sometimes in HubSpot. And so yeah, the best way is to ask the user because they will come— very few people actually click on a citation or on a source when they’re researching on ChatGPT. Instead, they’ll learn the brand name, maybe do a bit more research, and then type it into Google.
[11:04] Like, think of how you use it. That’s probably how you use it as well. So yeah, that is one of our case studies. And again, it’s in the article and there’s more detail. You can see the graphs and you can see examples of the articles we wrote and everything. But yeah, that was one of our big success stories. We were like, this is great, this is a framework that works.
[11:17] David Bain: It’s funny that how did you hear about us, it’s such an old-fashioned marketing way of doing things is actually so relevant now in the digital world.
Araminta Robertson: That’s right, I know, it’s very— yeah, it’s still our favourite way to track, to be honest, because it’s so complicated now with messy attribution.
[11:37] David Bain: So how important is this? And in the world of digital marketing now, maybe compared with conventional Google search nowadays? I mean, what percentage of someone’s awareness and traffic should a brand actually hope to get from doing this?
Araminta Robertson: Well, I think it’s huge. We have clients where 30-40% of their inbound leads come via LLMs.
[11:58] So this is not just one little channel, this is clearly like a big acquisition channel. What I do always say is that SEO— like, we still do SEO for all our clients— this, you have to do both together, SEO and GPT articles. They’re very interlinked.
[12:10] So I’m not saying don’t do SEO or stop that or replace that. It’s just because the attribution and the buying journey is so much more messy. Now, you don’t know if someone has found you via LLMs, via Google, you know, it’s really hard to truly know these things. And it’s very likely that they’ve probably checked both.
[12:31] I think there was a statistic the other day that I read in a study that like 70% of people check both. Google and LLMs. So what you want really is to be, I think, appearing in both. And I think doing SEO work will get you maybe halfway there, but we have found that doing this in addition makes a big difference.
[12:49] So it’s not, not just do SEO, but also do SEO plus a few little things that you can do for LLM visibility is really how you can maximise it.
David Bain: And you say that many different brands have had a success, you know, 30-40% of traffic or inquiries coming through this. But is that certain industry types? Are there certain industries that should absolutely focus on and others that shouldn’t?
[13:10] Araminta Robertson: I mean, it’s hard for me to say because we only work in fintech and financial services, but from what I’ve heard, it is across multiple industries. So definitely SaaS. I think any industry that has done well in SEO where they acquire a lot of their customers via SEO, I think 30 to 40% of their inbounds probably also come now via LLMs.
[13:29] But if you were like in an industry where it’s all trade shows, they probably are not as impacted. So it depends a little bit on that industry.
David Bain: And how future-proof is this strategy? If you— if an LLM becomes less popular, or there’s a future LLM that relies on an entirely different way of actually deciding what information to share, is it likely that what you’ve done in the past isn’t going to be useful for you in the future?
[14:00] Araminta Robertson: I mean, that’s hard to say, right? I cannot predict the future. All I can say is that we have to adapt. So this is a methodology that has come out of the need to adapt. I would say that there’s just no downside. Like, you know, this has to be still high-quality content that you’re creating. All you’re doing is more content.
[14:10] So, and in fact, we found that a lot of these GPT articles rank as well organically, and people will find them organically and convert. So in the end, what you’re doing is also helping with SEO. So far, it still works very well. In a few years, maybe, I don’t know, LLMs do completely change their model, but I also just don’t see— the only thing that could impact it a lot is if LLMs suddenly decide to not scrape Google, which I don’t see that happening because they need to.
[14:39] Google has decades of experience indexing information. It would— it only makes sense for LLMs to scrape Google. There is no other kind of indexing tool that is as good as them, so I don’t see that changing. The only, the only change is maybe if the LLM decides to like, yeah, basically not source from Google and only be closed, but then that would not be a very good model. So I don’t think so. For the moment.
[15:03] David Bain: Find out more and download your free copy of the Pump and Funnel Quick Start Guide over at pumpandfunnel.com. We’ll see you at pumpandfunnel.com. Now it’s time for the second part of the show, delving into Araminta’s marketing journey to date and uncovering her top marketing tips, starting off with work window.
[15:21] Araminta, can you give us an insight into your current working life? What differentiates the way you operate from others?
Araminta Robertson: Yeah, this was an interesting question to think about. I think, I’m not sure, to be honest. I think what differentiates us maybe as an agency is that we’re always willing to try new things and push the boundaries.
[15:44] So when there’s a new development, AI, whatever it is, we’re like constantly testing, constantly trying. We’re not afraid to disrupt our own model if necessary. So we’ve built a bunch of AI tools. We have our own AI writing tool. So I think, yeah, we’re, we’re never— we don’t procrastinate.
[16:03] Maybe I think that’s the way that we like to think about it, is like if there’s something scary to do, something that, you know, we know needs doing, we do it. We don’t wait on it. And I think maybe that differentiates us and that helps us.
David Bain: I wish I would have— I’d like you to look back on the early stages of your marketing career. What didn’t you do so well? What would you wish that you would have done differently?
[16:27] Araminta Robertson: I think the more I spend in my career, the more I realise how important in-person connections are in relationships. So I think maybe I wish I had started hosting events and building in-person relationships sooner, and that might have helped by moving to a city where a lot of that happens, like London, for example. So yeah, I think working on in-person relationships more. That would have helped.
[16:53] David Bain: Fast Track to the Future. Staying with your early working life, what’s one thing you did in the past that’s fast-tracked your future marketing success?
Araminta Robertson: Well, starting to host in-person events really sped things up. So once I started doing those, that really sped things up because I was able to build relationships and also have a much better insight into the pain points and issues that people had.
[17:11] So I’m really glad I started doing that, and I just wish I had started sooner, I guess.
David Bain: The This or That round. This is the quick response round. 10 quick questions, just 2 rules here. Try not to think about the answer too much, and you’re only allowed to say the word both on one occasion, so use it wisely. Are you ready?
Araminta Robertson: Okay. Yeah.
David Bain: Annual strategy or quarterly strategy?
Araminta Robertson: Quarterly.
David Bain: In-house or agency?
[17:40] Araminta Robertson: That depends on so much. I don’t know, both.
David Bain: Single channel or multi-channel?
Araminta Robertson: Multi.
David Bain: Paid ads or SEO?
Araminta Robertson: SEO.
David Bain: Email marketing or social engagement?
Araminta Robertson: What is social engagement? Social media?
David Bain: Yeah, interacting with social media.
Araminta Robertson: Social.
David Bain: LinkedIn or X?
Araminta Robertson: LinkedIn.
David Bain: X or X’d off?
Araminta Robertson: For our own marketing, no, no X.
David Bain: Human or automation?
Araminta Robertson: Human.
David Bain: Raw content or polished content?
[18:22] Araminta Robertson: Depends on a lot, but I would say maybe raw.
David Bain: Data-driven or creativity first?
Araminta Robertson: Data-driven.
David Bain: Got there without using multiple boths. A bit of a struggle. I mean, initially it was the— yeah, in-house or agency is the both.
[18:48] And then I think the human automation was the bigger pause after that in terms of both. Are you doing both as well?
Araminta Robertson: So much. Yeah, yeah. Oh yeah, I didn’t realise you could say both. You can only say both once. And what was the other one you could say once?
David Bain: Well, yeah, but both once, both is the one you can say once. Yeah.
Araminta Robertson: Okay. Okay. So I only said it once. That’s okay then. Exactly.
David Bain: Yeah.
Araminta Robertson: Tough, tough.
David Bain: But it was a struggle. There was, there was another one coming. That’s why I was just wondering about that. Okay.
Araminta Robertson: Yeah, yeah, definitely.
[19:21] David Bain: Let us move on to secret software. Share a lesser-known martech tool that’s bringing you a lot of value at the moment and why that tool is important to you.
Araminta Robertson: Yeah. So in the LLM visibility world, if you’re deep in it, this is a well-known tool, but if you’re not, then you might not have heard of it, but they’re Peek.
[19:36] And although there are a lot of LLM visibility tools out there, I have been constantly impressed with— I love their UX. While a lot of other tools keep adding new features, they keep it relatively simple. And I love how they’re constantly innovating. I like the guys as well, the founders. So yeah, Peek is my choice there.
[19:54] David Bain: And moving on from something that you currently use, something that you’re going to use, which is next on the list. What’s one marketing activity or tool that you haven’t tried yet but you want to test soon?
Araminta Robertson: I would love to test creating videos, quite simple business videos, and use those to rank on Google and LLMs.
[20:14] Ideally, these would be kind of like you interview an expert on the client’s team and then maybe put some animations together, quite a simple business video explaining a topic, complicated technical topic, and then yeah, publish that on YouTube. I’d love to try that.
David Bain: Great, great thing to try. Are you seeing more and more video results on LLMs as well?
[20:33] Araminta Robertson: For certain topics where it makes sense, yes. I think there’s just a big opportunity in dry business— like, YouTube is saturated for entertainment, food, music, you know, all the fun topics, but very few like ‘what is transaction monitoring’ type topics. Not many YouTube videos.
[20:55] So I think there’s an opportunity there to create video content on the dry business topics, but then use that to influence SEO, Google, and LLMs.
David Bain: Right, yeah, that answer also reminds me of a book called They Ask, You Answer by Marcus Sheridan. Yeah, great book. Yeah, absolutely. I interviewed him a couple of years ago.
[21:09] I knew— I know he presented in Edinburgh a few years ago as well. I’m not sure if you’ve really— you’ve seen him speak at all, but yeah, maybe not. But it’s a wonderful, yeah, piece of advice of just interviewing the top salespeople or the people that know a lot about your customers and your business, maybe 5 different questions or so, and just getting that essential content as part of your process towards actually driving new business.
Araminta Robertson: Yeah, yeah, exactly.
[21:37] David Bain: Well, let’s credit a couple of people who have helped you along the way. Now, Magical Marketers, name 2 marketers who’ve inspired the way you do things. Firstly, an established marketer, and secondly, an up-and-coming marketer. What can we learn from them and where can we find them?
Araminta Robertson: So my established marketer is April Dunford. I love her work. I think she has very unique, different helpful ideas.
[22:06] So her book Obviously Awesome was really helpful for us as an agency in positioning our own work and our own services. So I think you can just find her anywhere. She has a podcast, she has the book Obviously Awesome, she’s on LinkedIn. I recommend all the content that she creates.
[22:20] Up-and-coming marketer, I have picked Louise Linehan. She is on the marketing team at Ahrefs. And she’s just been creating lately a lot of really great content on the Ahrefs blog, which is very interesting, data-driven, some cool studies. I met her many years ago at Brighton SEO. And it’s just really cool to see her journey. And I think she’s someone who’s really worth following on LinkedIn.
[22:45] David Bain: My number 1 tip. I remember what’s the number 1 tip you shared in the Pump and Funnel marketing show.
Araminta Robertson: I would say create more content that is bottom of the funnel. So targeting people who are already aware they have a problem and actively looking for a solution, whatever it is, just create more content targeting those people. And that’s where you’ll see results across both LLMs and SEO.
[23:08] David Bain: I’ve been your host, David Bain. Araminta Robertson is founder of Mint Studios. Find out more over at mintcopywritingstudios.com. And grab your free copy of the Pump and Funnel Quick Start Guide over at pumpandfunnel.com. Until we meet again, stay hungry, stay foolish, and stay subscribed.
Further Reading
- The CMO’s 2026 Roadmap: How to Turn LLM Visibility Into GTM Strategy
This article analyzes how brands and competitors appear across ChatGPT, Gemini, and other LLMs and explains how to use LLM audits to identify which media, communities, and partnerships most influence AI-driven discovery, directly complementing the episode’s discussion of targeting specific LLMs and measuring visibility. - How to Get Your Brand Cited in ChatGPT (Verified Tactics for 2026)
This guide provides verified optimization tactics including using comparison tables (which the episode mentions as important), structuring content with FAQs, and updating content frequently, aligning with the GPT article framework’s emphasis on formatting and freshness. - What is Generative Engine Optimization (GEO)? 2026 Guide
This comprehensive guide covers research-backed GEO strategies including comparative listicles and statistics integration, which directly supports the episode’s recommendations for product-focused comparison articles and factual content.
