Don’t get me wrong: I love AI and I’ve been an avid user of various chatbots and such since the very (publicly available) dawn of it. I guess, it’s been over 4 years of me using AI almost daily for a massive number of tasks, from work-specific ones (I’m a product marketing manager in tech, working with B2B SaaS products for 4+ years) like outlining a go-to-market strategy and being my brainstorming buddy to personal stuff such as serving as my therapist.

Anyway, I don’t think the problem is AI. I think that the problem is that we as humans are so lazy that we will happily outsource the most energy-consuming task for $20 a month.

Yeah, thinking. Critical and creative.

Have you wondered why all articles now sound just plain? Boring? Like there’s something lacking, maybe a general idea that must be running through good writing and an author having their own discernible opinion? Je ne sais quoi.

The same goes for any product and brand out there. Vibe coding is great but have you ever tried it yourself? Without some extensive prompting exercises, you won’t get far, trust me, because everything that can be produced quickly is nearly the same. Why? Because all AI models are in constant training and they are trained on virtually the same data that is open-sourced and most of open-source data is AI slop.

My point is, the output we are getting from chatbots and AI agents is universalized and generalized and that’s how we get to the point of looking all the same. AI is extraordinarily good at helping us answer questions.

But when you outsource the question itself, congrats, you’re in trouble.

Dealing with faceless articles

See, copywriters and ghostwriters have been around for ages. I get it. That’s fine. I’m not asking for the highest quality of content because not every business or product needs it. Fact-checking is great, but it’s not the biggest of my concerns either.

Like I’ve mentioned before, AI generated content just doesn’t say anything. You read an article that is grammatically correct, filled with facts backed up by links, probably even with the right interpretation. It’s good but when you finish it, you are left with a feeling like you’ve been robbed of your time.

So, the number one thing is storytelling. We as humans naturally gravitate towards a good story. That’s why we love chatting and gossiping so much whether it’s a family gathering when your grandma is telling a story of how she went abroad every summer “to fix her health” (which means have a summer fling) (based on true events, by the way) or just snickering with a friend because your mutual enemy had disgraced themselves at someone’s party.

What makes storytelling good? First, connecting with your audience on an emotional level. AI can never guess which tiny, highly specific emotional triggers to use. Because it is human, and only we can stay within the context of your audience’s situation and problems and know what is really worrying them and why. Take politics for example. Me and you know that the prices are going up due to certain events, and we both probably have someone in mind to blame. It’s not rational but it’s emotional, right? We have sympathy, and we are guided by our emotions. Such a beautiful humane feature, isn’t it?

Second thing is personality. I bet you’ve noticed those little hints of my background information I am sprinkling here and there - at least you now know what my marketing experience is and what my late grandmother did every summer. Do you feel just a little bit more connected to me? Because I hooked you with my own opinion at the beginning of an article and then continued sharing a little more throughout. You feel like a part of someone’s story.

And the third thing - a great article/story/post/whatever offers a takeaway, a lesson, or a fresh perspective that makes the narrative feel like a valuable gift. That’s what I meant when I said that AI generated texts are lacking a general idea. Maybe in the future AI will do it better but now… we need to think about that main idea in advance.

Branding

So, what makes a brand great? Let’s think of Apple, for example. I grew up in 00s and 10s so I remember the iPhone craze. People either desperately wanted one, proudly had one, or passionately explained why they would never buy one. Here’s why it stirred up so many reactions, both positive and negative:

• a distinctive design. The product was doing part of the branding work because there was a product philosophy behind it
• killer features that none of the competitors had
• Mr. Jobs (you can like him or hate him but you won’t say you don’t remember the way he presented new devices or dressed, will you?). Sometimes the easiest way to make a brand human is to let a human being be present
• an inspiring tone of voice, think of that commercial “Think different” which isn’t even grammatically correct. My fifteen-year-old self could cry listening to, that’s how inspiring it was, selling the idea of who I can be!

Personality and distinctive branding come from having preferences, and the best news is that we all have them already. Just let them out, fellow PMMs. Be less strict with your TOV. Or visuals. Ignite some thought or a conflict inside your customers’ heads.

Imagine asking an AI to create the next “Think Different”. It could probably give you fifty slogans about innovation, creativity, challenging conventions and empowering people. All of them would sound fine, some might even be very good. But would any of them feel inevitable in retrospect? Would you know that this sentence had to come from this particular company, at this particular moment, with this particular belief about the world?

So what can we, as product marketers, actually do about this?

1. Decide what you believe before asking AI what to write

This sounds painfully obvious, but I think it is one of the most important habits we can develop.

Before opening ChatGPT, Claude, Gemini or whatever your flavor of the month is, write down the answer to three questions:

What do we believe about this market?

What do we believe that our competitors don't?

What do we want our audience to believe after reading our article/post/looking at our website? After interacting with our product, after going through onboarding?

Then AI becomes incredibly useful. Give it your point of view and ask it to challenge your assumptions or your argument and point out where it is weak. I frequently use it to generate counterarguments, alternative framings, examples or ways to explain the same idea to different audiences.

2. Build a point-of-view based brand book

I’d suggest creating an opinion book. Or a point-of-view book, whichever you prefer. Gather all of your ideas and opinions and your brand mission and build it within your prompting.

You can put things in ChatGPT’s memory (with certain wording) like:

• our strongest opinions about the market
• assumptions we disagree with
• industry clichés we hate
• things our customers consistently misunderstand
• phrases customers actually use
• phrases we never use
• claims we can prove
• claims we deliberately avoid
• examples of campaigns and content that feel unmistakably “us”
• examples of things that were technically good but felt completely generic
• topics where we are willing to be provocative

I’d go even further and add the reasoning behind some of those choices.

3. Humans should stay in charge

I don't think every sentence needs a human sitting there and lovingly crafting it with a cup of coffee. There are better uses of our time, yet every part of your marketing kit should have a human owner, whether it’s a design concept or an article. And that human owner should carry an idea. And be responsible for the argument, evidence and final judgment.

And please - please - ask yourself a question “Could this have been created by any competent marketer with access to ChatGPT?”. Yep, not everything needs to become a piece of groundbreaking thought leadership, but don't confuse usefulness with differentiation.

4. Give your experts a microphone

What makes your marketing human if not humans, right? People love people. And people love other people’s stories and cases and ideas and solutions to their own problems.

A strong expert-led article is interesting not because an expert's name is attached to it, but because they have actually seen something - well, they have a story to tell because they FAILED. And they are imperfect. So are we all.

I want us to use it much more aggressively but lower down the stack.

Use it for the work that is expensive yet repetitive: research synthesis, transcription, analysis, first drafts, repurposing, localization, competitive monitoring, content operations, formatting, testing. Basically, all the things where the value comes from speed, scale and consistency.

Use it as a thinking partner for work that benefits from challenge: testing hypotheses, exploring positioning alternatives, finding objections, scenario planning, segmentation, GTM planning.

And may you stay beautifully human.