Hi everyone,
Denisa here.
I’ve been a content strategist at Mint for a bit over a year now, and as Elliot is getting married this month (🎉), I’ll be taking over for this newsletter.
(Here’s a bit more about me and why I joined Mint.)
Today, we’ll discuss something I’ve been thinking about a lot lately: how much of the content creation process should we hand over to AI?
On one hand, you've got leadership pressuring marketing teams to use AI more efficiently.
On the other hand, more and more buyers are protesting and critisizing brands that overuse AI in their designs, content and communications.
So where's the balance? Where does it make sense to use AI, and where does it make sense to avoid it?
After all, just because AI can do something doesn’t mean it should.
Over the past year, we’ve been experimenting with AI workflows and automations at Mint, and what I’ve found is that human insight is still very much central to content marketing. While AI can speed things up, the trick isn’t to use it everywhere; it’s to use it where it actually helps you produce more high-quality content.
So, here’s when it makes sense to use AI vs. human insight, creativity, and judgment.
The risks of using too much AI in content marketing
Using AI to help write an article about embedded payments or stablecoins isn’t necessarily a problem. The problem is when companies use AI to produce enormous amounts of low-quality content on a low budget.
What they often don’t realize is that creating more content doesn’t necessarily mean getting better results.
In fact, if you mass-produce AI content, you risk ending up with articles that:
- Don’t resonate with your readers because they’re generic, poorly written, or don’t really offer anything new. Often, your website content is the first touch point a prospect has with your brand. Mass AI-generated content doesn’t really give the best first impression.
- Bring in irrelevant traffic that doesn’t turn into revenue because the articles are not built around a clear strategy or an understanding of the ideal customer
- Can even describe your product inaccurately, potentially leading to compliance risks (which is especially important in fintech)
What’s more, this type of strategy can get you in real trouble with Google.
Google doesn’t have any issues with using AI to produce content – in fact, based on a study of 100,000 random keywords, only 13.5% of the top 20 ranking pages are "pure human”, and 81.9% include some form of AI assistance.
But as Elliot described in his last newsletter, it is flooding the SERP with mass-produced, low-quality and unhelpful AI content (also called “Mount AI”) that is the problem. As a result of this, Google can crack down on your website, as seen in this example of keywords increasing in visibility before dropping down quite quickly when Google realized what was going on:

What’s more, the most recent Google spam update in September 2026 specifically targeted mass AI-generated content – which clearly shows these types of articles are classified as spam.
The way I see it, this initiative from Google’s side is actually a huge sign that things aren’t all over for good content. When it comes down to it, Google still cares about quality first and foremost.
And the people using Google to search for information care, too. Just think about what you appreciate about human-made content: it has personality and unique takes that you wouldn’t find elsewhere.
It’s no wonder that Reddit, a forum based on personal experiences and opinions, has more than doubled in daily active users in three years: the site went from 60.4M users in Q2 2023 to 130.3M in Q2 2026.
This is much quicker growth than before – and remember that ChatGPT first rolled out toward the end of 2022. It could very well be the case that the more AI content there is online, the more we crave human insight.
But this isn’t to say AI is not a useful tool. It can be, but it’s how you use it that matters.
Two examples of where to use AI in content production
1) Scaling content (carefully): Use human insight for the input and AI to speed up the output
The quality of your AI output will only be as high as the quality of your input.
If you ask ChatGPT to write an article on embedded finance, you’ll likely end up with a generic piece that’s just a rehash of what’s already available online.
If you give it an interview with an expert, a comprehensive outline based on keyword and topic research, and a bank of information about your company and products, though… the end result will be a very different story.
This is because the input is what humans are best at in content marketing: deciding what topics are even worth writing about, researching competitors, conducting interviews, reading customer reviews, and having a sense of what information can be genuinely useful.
AI, on the other hand, is good at quick outputs: turning those insights into a draft that would take hours or even days to type out.
At Mint, we’ve arrived at a workflow that’s the best of both worlds: human judgment continues to play a central role while AI speeds things up.
First of all, we target very specific bottom-of-the-funnel keywords that bring in readers in the late stages of the buying journey. You can read more about that here. This means we put a lot of critical thinking into the keywords we pick in the first place.
Then we write up to X pieces of content per month, with half of them still being written by a person. The other half goes through an AI-assisted workflow, which looks like this:
- Human research and interviews: A content strategist interviews an in-house expert, and the writer researches the topic and prepares an outline based on the interview.
- Human review: The strategist checks the outline for accuracy, completeness, and logical flow.
- AI-assisted drafting: The writer uses our internal tool, Plume, to draft the article. Plume also draws on previous articles, interviews, product pages, and client notes to suggest relevant information and quotes.
- A mix of human and AI final edits: The strategist reviews the draft, the writer implements feedback and uses an internal fact-checking AI tool. The strategist then collaborates with the client on final revisions.
This allows us to create more content than we could pre-AI, while still maintaining the same quality.
2) Behind-the-scenes workflows: Use AI for repetitive tasks and human insight for strategic thinking
Internal workflows are where I’ve found AI to be most useful.
As a content strategist, a lot of my work revolves around reviewing content, keyword strategy, troubleshooting, client communication, or analyzing and reporting on data. And a lot of this requires critical thinking and a human touch.
But there’s also a lot of repetitive work that goes on behind the scenes, and by automating these smaller tasks, I've been able to spend less time on admin and more time on strategy.
For example, we’ve been experimenting with using AI to:
- Automatically create and assign briefs to writers
- Update internal documents with the relevant client information based on updates or calls
- Interlink between articles
- Fact-check information from third-party sources (after the writer has done so manually)
- Summarize calls with clients and outline monthly results emails based on that information
- Double-check interviews with experts to ensure we’ve covered all important angles in articles
- Create reports to visually showcase data from SEMrush, GA4, or GSC dashboards
- Automatically update articles for higher rankings or consistent product messaging
Here’s an example video of the last point in practice, which helped us scale from optimizing 1-2 articles per month to 6-10 per month.
Read more: How we can help put AI to work in your content engine
Of course, each task still requires your thinking in the end – if Claude puts together a report on why your website’s organic traffic has changed, you still need to understand why those changes happened, what they mean for your company, and what to do next.
But a lot of the technical part is handled much more quickly and smoothly, which means you can invest all the more time into what you’re best at as a human: analyzing and solving problems.
So, where should you start with AI in content?
If you’re thinking about introducing more AI into your content marketing process, I’d start by asking yourself:
- Which repetitive tasks could we automate with AI so that we have more time to do more valuable work? E.g. automatically detecting and notifying you of which pieces of content have lost visibility in the last 30 days.
- Which parts of our work genuinely benefit from our experience, creativity, and critical thinking?
- And where can AI help us produce content faster without sacrificing quality?
One last thing to remember is that the goal shouldn’t be using AI for the sake of it. Just because you create more content or implement more automations doesn’t mean you’ll get better results.
What’s more, overusing AI can quickly backfire, and it might also turn into a workflow that feels inauthentic and, to be honest, unenjoyable. It’s often understated how much enjoyment matters in content creation – it’s what gives it that spark that makes for a better piece.
The goal is to use AI in a way that grants you more time – not less – to do joyful, valuable, human work.
Mint Studios Recommended Reads
- Google is cracking down on AI slop. Is your content safe? A previous newsletter about improving or removing AI content to stay aligned with Google’s policies
- Stop using AI for writing. An interesting piece by Kieran Flanagan, HubSpot's SVP for Agentic GTM & Systems, about the importance of joy in content creation
- How we use AI for every article without making AI slop. An Ahrefs article that gives some more interesting tips on implementing AI in your content creation process, like using Wispr Flow (we do this, too) to let AI interview you in the brainstorming stage
Thanks for reading,
Denisa & the Mint team 🎉
We help financial services and fintech companies turn content into measurable pipeline. Learn more about what we do.










