I was an AI evangelist. Now I’m AI vegan. Here’s why.

It’s been a while since I’ve written to you.

Lots of people have joined here since my first ‘AI vegan’ story, and I wanted to elaborate on the whole thing in more detail. Voila.

Running a LinkedIn content agency showed me how generative AI is sucking the humanity out of social media, personal branding and the stories only we can tell.

Do you recognise this feeling? I’m scrolling through my LinkedIn feed, skimming self-promotional anecdotes and the latest rage bait, before I find myself hitting “see more” to read a full post. But then something feels off.

It’s uncanny. The words are all there, in the right order, trying so hard to sound profound. 

Short, choppy sentences.

“It's not about content. It's about connection.”

“Because you deserve to be listened to, not just heard.”

It has the whiff of AI slop, the synthetic “thought leadership” content that’s become the norm of LinkedInfluencing. It’s everywhere. If you took away the profile picture, the headline and the person’s name, you wouldn’t have the faintest clue who wrote it.

For a moment, you think a post like that is legit. Perhaps the hook pulled you in. It almost feels like it’s saying something of substance. It looks like content but tastes like cardboard. It looks like writing yet there’s little original thinking there. The more you read it, the less it says.

Until recently, I felt like AI slop was atrophying my brain and disconnecting me from my craft. 

I knew I needed to quit. But that’s not so easy when your whole business is built on AI.

I faced a choice: keep eating slop or go “AI vegan”

Confession: I used to be an AI evangelist of the highest order. I was running a growing LinkedIn content agency, creating thought leadership for founders, executives and entrepreneurs. I take pride in my work, connecting with clients’ stories. But I also felt the pressure to scale, so, in December 2024, I built a suite of custom GPTs to handle 80% of content drafting. I followed the 10-80-10 rule: 10% human prompting, 80% AI drafting, 10% human polish before publishing.

I’m not going to lie: at first, this seemed like a harmless way to grow the business. AI felt like the golden goose. I'd trained my custom GPTs on hundreds of thousands of words I'd written by hand since 2022. With detailed briefs and knowledge bases trained on my clients’ tone, writing style, business and personality, my chatbot army did a scarily good job at imitating my clients' voices. I’d uploaded transcripts from hours and hours of interviews with them, too. 

But over the early months of 2025, I found myself in a strange situation: I stopped evaluating and editing content based on its resonance, relevance and interest to the target audience. Instead, I focused on trying to assess whether the content would pass a different test: would the client think a human wrote it? If I felt the content met that standard, we shipped it. We adapted our GPTs’ instructions to exclude classic AI giveaways like “it’s not just X, it’s Y.” Fluff like “supercharge,” “revolutionize,” “fast-paced” and — most tellingly of all, if you believe the “experts” who’ve been copywriters for 27 minutes — we trained it not to use em dashes. 

As a result of this efficiency and optimisation, we shipped more content, faster than ever. All of a sudden, the barriers to scaling my content agency had all but disappeared. The sky was the limit. Or so it seemed for a short while. 

Then one day, something shifted. I was reading a fresh batch of AI-generated content we’d created for a client. And it hit me that this was exactly like the slop I hated reading in my feed. I was so focused on “humanizing” synthetic drivel that I’d lost sight of the humanity in the work. 

That’s when I knew I had to go “AI vegan,” by which I mean, I had to become a human writer again. It sounds almost silly to say that, but just as many people make dietary choices to improve their health and reflect their values, I knew I had to find my way back to what was real. 

AI vegan wasn’t merely my new marketing position. It was an awakening that prompted me to rebuild my business around human skills. And to do that, I had to reconnect with the craft of storytelling.

Wes Anderson (unknowingly) wrote my manifesto

I help clients show their authentic selves, share things of real value, and spark conversations that real people want to be a part of. That’s my craft. But I’ve found it’s a lot harder to do those things for yourself. And, paradoxically, sometimes one of the best ways to reconnect with your own craft is to hear the greats in other disciplines talk about theirs. For me, that happened to be Wes Anderson. 

Yes, he has one of the most instantly recognisable styles in all of cinema. It’s not just look and feel. It’s casting, script, dialogue, set design and cinematography. How does he capture this unique identity, over and over again? Anderson recently told Vanity Fair about the process across his 12 films:

I never expected Anderson to give me a manifesto for an “AI vegan” LinkedIn content agency, but this message instinctively rings true for my craft, too. 

My favourite part of the creative process is digging into a client’s life story and seeing what’s underneath the surface. This is why all my work starts with an interview. If you want to start conversations and stand out (and no one could argue that Anderson’s work blends in), put down that shiny new ChatGPT prompt and pick up a shovel. 

Your best ideas, your best stories and ultimately your best content, are there to be uncovered, not prompted into existence. Get digging into your calls, conversations, your journal notes and scribbles. This is where you’ll find stuff that will make other people feel something. 

The antidote to slop, AI or not, is knowing yourself

For me, the digging starts by surveying the field — getting into the client’s world, understanding what they do and who they do it for. 

But we really make progress in an interview. That’s where I create the space for the subject to share what they know. This inevitably means long pauses, ums and ahs, lots of gazes out the window. 

Most clients I interview are unconsciously competent, by which I mean they can’t always describe their expertise; they just know what to do. The interview process is a powerful step to getting them thinking clearly about the value they bring to the world. 

So I’ll ask lots of “what” and “how” questions about the people, places, objects and environments in your life that shaped you. Questions like: 

• What do people come to you for advice with most often?

• What do you notice that no-one else sees?

• Think about something that brings you joy … what do you love about it?

Those sound deceptively simple, but they can really stump you. When I first went AI vegan in June, I found I was circling these questions myself. 

I understand why slop proliferates. Sometimes, making slop just seems easier. And just like an army of custom GPTs can generate a bottomless supply of it, humans have been posting slop on the internet long before large language models came into the popular lexicon. But if you want to say something of meaning, if you want to stand out, if you want to be distinct, slop won’t do.

I’d argue that the antidote to slop, AI or not, is knowing yourself. And that can take quite a lot of excavation. It’s the only way to uncover anything of value.

Almost two months into going AI vegan and I'm straining storytelling muscles I’d forgotten about, writing more sharply and building my business around what only humans can do.

The work can feel harder, but it’s mine again.