Fintech Content Expert: How to Turn Content into Pipeline

Araminta Robertson built Mint Studios around a hard content problem: proving that fintech content creates pipeline. As Managing Director of a content agency for fintech and financial services companies, she works in a category where a weak article is not just forgettable. It can damage trust when buyers are evaluating money movement, compliance and technical risk.

Her starting point was not an agency playbook. At 18, she started a personal finance blog in Barcelona after realising she knew nothing about taxes, mortgages or investing, then turned that curiosity into a fintech specialism spanning clients such as Persona, GBG and ClearBank.

This conversation gets practical on AEO because Araminta is clear about what she has tested, what she has not proved, and what she would avoid. Her GPT framework starts at the bottom of the funnel, uses short expert-led articles to answer specific prompts, and treats AI as an assistant to judgement rather than a replacement for it.

This interview has been edited for clarity and concision.

Tom: What made fintech and financial services the category you wanted to build Mint Studios around?

Araminta: I call myself Managing Director rather than CEO because we do not have a C-suite. Mint Studios is a content marketing agency for fintech and financial services, and we help those companies turn content into measurable pipeline. We have worked with companies like Persona in the US, GBG and ClearBank.

The backstory starts when I was 18. I had finished high school and realised I did not know anything about finance, which is the case for a lot of 18 year olds because we are not taught about taxes, mortgages or investing. I thought that if I did learn those things, it would set me up well for the future, so I started a blog.

I was living in Barcelona, where I grew up, and I would read personal finance books like The Intelligent Investor and Rich Dad Poor Dad, then write about them for other young adults. It was a “let’s figure this out together” kind of blog. I met a fintech company at a business event, they saw the blog, liked it, and hired me as their content marketing person.

I loved how learning about personal finance gives you control because you start to understand how the world works. Now we do more B2B than personal finance, but payments and finance are like the arteries of the world. Without them, the world does not work.

Tom: Why is fintech content harder to get right than content in a lower-risk category?

Araminta: The stakes are higher. Opening a social media account or a SaaS tool account is not the same as opening a business bank account or choosing a banking-as-a-service partner. From an SEO perspective, that puts you in the YMYL bucket, which means your money or your life.

Google uses that category for topics like finance and healthcare because bad advice can have serious consequences. Telling someone to invest in the wrong crypto coin and lose their life savings is very different from recommending a $5 a month tool. Your content needs to be more rigorous and more careful.

It is also more complex. Writing about payment infrastructure is harder than writing a recipe. You need strong writers who can make that complexity easy to understand.

The last piece is expertise and trust. In B2B fintech, people want advanced and technical content, but they also want it to come from someone who knows what they are talking about. Trust matters more in financial services because the stakes are higher.

Tom: You have said expert interviews are non-negotiable. What do they add that desk research and AI cannot?

Araminta: When we first start working with a client, we spend a lot of time interviewing the experts on their team. Sometimes that is for each article at the beginning. Later, as topics repeat and we understand the product better, that becomes less necessary.

Those interviews help us understand the product, the pain points, the right topics, and the reader’s level of knowledge. The content can then be as complex as necessary and still be clear. When a client says, “wow, this is really good”, the answer is usually that it is because we interviewed an expert.

A few years ago, clients sometimes pushed back and asked if we could just do desk research. We no longer get that pushback, and I think that is because of AI. People can see that if you do not interview an expert, the content risks becoming AI slop. If you are saying the same thing as everyone else, why should anyone read it from you?

There is still a place for AI, but I always say AI-assisted rather than AI-generated. Your output is only as good as your input. If a client already has a strong body of content, AI-assisted work can be very useful. If they have little content, or the existing content is not strong, a human writer needs to be much more involved.

Tom: How does your GPT framework decide which AEO content is worth creating?

Araminta: Around September or October 2025, we developed a framework to help companies influence LLM visibility and get recommended more often. We call it the GPT framework.

The whole thesis is to start at the bottom of the funnel. These are people who know they have a problem, are actively looking for a solution, and are much closer to converting. We take the same approach with SEO and content: start there, then work your way up the funnel.

First, you decide the bottom-of-funnel topics. Then you put together prompts that people might be typing into ChatGPT, Gemini or other tools. For one topic, you might create 10 prompts. You will not get every variation exactly right, so you need to understand the audience and put yourself in their shoes.

Then you create short articles, usually 500 to 800 words, answering each prompt. I think of it like a hub-and-spoke model. In the centre, you have an in-depth SEO piece, for example on open banking APIs. Around it, you create shorter GPT articles answering specific questions, such as an open banking API that does CVRP or one that works across Europe and not just the UK.

The key is choosing the right prompts. That is 80% of the work. If you target something like “what is open banking?”, ChatGPT will probably give a straight answer and may not mention any brands. There is not much point targeting that if your goal is to be recommended.

Tom: Why not just add those prompt answers as FAQs under one big article?

Araminta: We have found that separate pieces do a better job than only adding FAQs. Each piece can still have FAQs, and we like adding them because there are so many ways someone can ask a question.

My theory is that when you answer the exact question on a separate page, with context, it is easier for the LLM to find and understand. The URL can match the question much more closely, and the answer is directly there. In an FAQ section, there is more surrounding information and the URL is often less specific.

Long URLs seem to help because they are more specific. When ChatGPT searches online, it can break the prompt into keywords and look for URLs that match. If your URL contains the words from the prompt, and your meta description answers the question clearly, you are giving it something easy to retrieve.

We track the prompts we target with tools such as Peec and Profound. Often, within days or even hours of publishing GPT articles, we see visibility shoot up. The next question is whether that brings conversions, and so far we have seen inbound increase when companies publish more GPT articles around the right topics.

Companies that are quite early stage, sometimes even too early for SEO, can benefit a lot from this because LLMs are faster than Google at showing brands. If you target one prompt, you can start appearing for related prompts too, even when the wording is not exactly the same.

Tom: Where do SEO and AEO support each other, and where do you think the advice gets lazy?

Araminta: GPT articles can be a useful testing ground for SEO. Sometimes we have a hunch that people looking for a certain topic might have the right intent, but we are not sure. We can publish shorter GPT articles, see whether they work and whether people start coming in, then decide if it deserves a full SEO article.

A lot of those topics show zero or NA search volume in Semrush, but people are still searching for them. I think part of the reason is that people are starting to use Google more like Gemini, with longer and more personalised searches. The long tail is getting longer.

The biggest issue in AEO advice is that a lot of it is vague. People say things like “do chunking”, “make your headers like this” or “make it more conversational”. That is not always wrong, but it is not very helpful. I like a repeatable framework that can be scaled over time.

I am against mass AI content. Publishing hundreds of AI-generated pages a month is a good way to get penalised by Google. The “rank to tank” approach, where people mass produce pages to rank quickly and then watch them fall, is risky.

There is also a lot of talk about Reddit, but I find it hard to manufacture because it either feels authentic or you struggle. I have tested LinkedIn articles and Medium articles and have not seen a big impact. Digital PR is good for the long term because it builds authority, but I have not found a repeatable framework that works as well as GPT articles.

Tom: What should fintech teams keep in-house, what should they expect from an agency, and what skills matter in content marketers now?

Araminta: We only take on clients where we do SEO content and AEO together. We would not do AEO only because it is such a small part of the work, and it is tied so closely to content production. If we are producing content, it makes sense to do SEO as well.

If a company already has a mature content programme and is doing well with SEO, it can implement the GPT framework internally. We had a client where the work reached a natural point, results were strong, and there was not much left for us to do. We passed the framework to them, and when we checked in later, LLMs were bringing in 20% to 30% of their inbound, with the best quality leads.

For an agency relationship to work, the first thing is trust. If a client wants to decide every topic and does not want to rely on our experience, that makes the work harder. They hired us to bring them customers, so they need to let us do that.

The second thing is access to subject matter experts. If the experts are too busy and we cannot speak to them, it does not make sense for us to work together. In the first two or three months, we need more access to get the product, pain points and tone right. After that, it tails off.

Strong positioning and clear USPs also make a big difference. If the positioning changes every month, it becomes hard to know what to say in the content or which keywords to target. More mature companies may change positioning once or twice a year, which is manageable.

When we hire, we mainly hire writers first and they may become strategists later. We look for clarity of thought, empathy for the reader and curiosity. It is no longer enough to be a great writer because companies also want to know how content turns into customers. That connection to revenue is more important now than ever.