AI’s Biggest Gotcha
And how to become the thing the machine can’t replicate.
Here’s a thought that should keep every GTM pro awake tonight.
Every time you prompt ChatGPT to write business-personalised copy, every time you iterate instructions in Claude, every time you talk to an LLM, you’re not just doing your job. You’re creating training data. And that training data is being used to build the thing that replaces you.
This is AI’s biggest gotcha. Not that robots are coming for your job. That you’re teaching them how to do it - enthusiastically, voluntarily, and at speed.
Do you see the mechanism here?
Step 1: Humans use AI tools to do GTM work - prospecting, copywriting, sequencing, personalisation.
Step 2: Every interaction generates high-signal data. Not social media noise or generic web scraping. Specialised knowledge work. The good stuff.
Step 3: That data trains better AI models.
Step 4: Better models need fewer humans.
Step 5: Go back to Step 1, but with fewer people in the loop.
It’s a feedback loop designed to consume its own creators. And the GTM-Clay-Ecosystem is ground zero because the data being generated isn’t low-signal fluff - it’s the precise, domain-specific intelligence that AI needs to replicate (and eventually outperform) the humans producing it.
The low bar makes it even worse.
Cold email outbound - the channel this entire ecosystem obsesses over - has set the bar so catastrophically low that AI doesn’t need to be brilliant. It just needs to be not terrible.
When the standard is spray-and-pray volume, generic “personalisation” scraped from LinkedIn, and messaging so interchangeable it could come from any company in any industry - AI clears that bar on day one. Tools like Octave and Twain already produce copy that converts as well as (or better than) the average human SDR’s output.
And with orchestration platforms like Clay and N8N handling the plumbing, very few humans are needed in the loop at all.
The US market pioneered this volume-over-quality approach to overcome embarrassingly low ROI. Now AI is perfecting it. And the people providing the training data - those who spend 24/7 building with AI, - will be the first to feel their own redundancy.
That’s the gotcha.
What makes this particularly vicious is that opting out isn’t really an option.
If your competitors adopt AI and you don’t, you fall behind. So you adopt it. Which generates more training data. Which makes AI better. Which displaces more humans. Which forces more adoption.
Not using AI becomes a competitive disadvantage - so you participate in the very loop that makes you redundant. It’s a race to the bottom where human labour is the first casualty, and the runners are building the track as they sprint.
Short term? AI-driven volume dominates. The low-bar equilibrium stabilises. Only AI-driven teams survive, with humans relegated to niche oversight roles.
That’s the operational reality.
But here’s the felt reality.
I feel it. You probably feel it too.
As a prospect on the receiving end of this machine, something is viscerally missing. As a consultant working with B2B brands, I see the missed opportunity at every touchpoint. The absence of brand differentiation is palpable - because the people who used to nurture it, the brand and marketing people, have been squeezed out. Or made responsible for the entire pipeline so their priorities shifted to metrics that reward volume at all costs.
There’s a latent demand building. A growing dissatisfaction with AI-generated noise that currently has no economic consequence - but will.
Because here’s what the automation-first crowd keeps forgetting: brand is the foundation for growth. Outbound performs better when there’s brand awareness, recognition, and trust behind it. Strip that away - which is exactly what’s happened - and you get diminishing returns dressed up as efficiency.
The market will saturate with AI-generated sameness. It’s already happening. My inbox is proof. Yours probably is too.
But there’s light on the horizon.
The Only Thing AI Can’t Automate
Brand.
Not brand as a logo or a colour palette. Brand as in the strategic intelligence that determines what’s worth saying, to whom, in what voice, creating what feeling. The human discernment that separates “this resonates” from “this is noise.”
AI can generate copy. It can personalise at scale. It can orchestrate sequences with ruthless efficiency.
But it can’t decide who you are. It can’t feel what your customer feels. It can’t build the positioning that makes someone choose you over a competitor who sounds identical.
The gotcha has a counter-move: become the thing the machine can’t replicate.
Brand strategists, messaging architects, people who understand customer psychology and emotional resonance - these aren’t legacy roles being made redundant. They’re the only roles that survive this loop intact. They’re the human-in-the-loop that gives AI something worth amplifying.
The rest? The workflow builders, the sequence optimisers, the volume-over-value operators?
They’re training data. Whether they know it yet or not.
So, Which Are You?
The person building brand equity that AI amplifies?
Or the person generating training data that makes your role disappear?
Because the loop doesn’t care about your career prospects. It just keeps spinning.
And it’s spinning faster every day.
Typos are my way of checking that you’re paying attention - or proof that my brain moves faster than my fingers. (Jury’s still out.)
