Is It Hypocritical to Sell Against AI Slop and Write With AI?

No, it isn't hypocritical to use AI to write sales content. A machine touching the words was never the problem. Half of us draft with it. The problem is when a machine is the only thing that touched the deal: no read of the human, no minute spent, just volume with a first-name token stapled on top.

I use AI in my own writing. I also sell a product built against AI slop. Both are true at the same time, and the people who pretend otherwise are usually the easiest to catch.

Is it hypocritical to sell against AI slop and write with AI?

It's only hypocritical if you pretend you don't do it. Here's my own version. One morning I used AI to help write something I published. An hour earlier I'd seen an obviously AI-generated brochure at a market and felt my whole face do the sneer. Same guy, same morning, both reactions completely sincere.

Everyone in B2B is quietly two-faced about AI right now, me included. We're producing the thing we resent, leaning on the tool we're a little scared of, and mostly not admitting either half out loud. Call it hypocrisy if you want. I think it's an identity crisis, and it's hitting the skeptic and the true believer at the same time.

If a model can draft the email, build the deck, and summarize and then take actions during and after the call, then the thing you were proud of being good at just got cheap. That's the flinch underneath a lot of the loudest anti-AI takes. I'll stick to my domain. Sales.

The thing agents can now do was never the thing that actually won the deal. Deals die inside the buying group, after the meeting, when a stakeholder got scared, asked for the thing that creates more indecision, and went quiet before anyone caught it. No bot flooding the top of the funnel touches that. The read of the specific human still decides it.

Why do the loudest AI critics write the worst AI content?

Because a lot of them never learned to drive it. The pattern goes like this. Someone screenshots a bad cold email, captions it "this is why AI is ruining sales," and a crowd claps. Then you look at how they actually use AI, and it's one-line ChatGPT prompts. "Write me a follow-up email." That's the whole relationship with the technology.

They're right that the output is garbage. They're wrong about why. The email is slop because the person driving it never learned to drive. A teenager stalling the car in a parking lot isn't proof cars don't work. We didn't ban planes because the first ones fell out of the sky, and we didn't ban the internet because 1999 was popups and Ask Jeeves.

I hate AI slop more than almost anyone. I built a company on hating it. But hating slop and being good at AI are different skills. One is a personality. The other is a practice.

What actually gives AI-written content away?

The template gives it away, whoever delivers it. I tested this on myself. I put out two clips of me, about 20 seconds each, and asked which one was AI. The real me read pure slop, straight-faced, every line: "Let me be honest with you. It's not about the product. Read that again. Let that sink in." The AI version of me read the sharp lines about reading the human in front of you and catching the second they start to freeze. One person guessed, and they got it right.

So the human said the machine words, the machine said the human words, and the one guess still landed. What tipped it was the writing. You can hear a template in about a second no matter who's reading it.

The platform ran its own version of the test. My best piece of content this year drew 6,789 impressions. The video carrying my AI twin drew 93, and 93% of those views stayed inside my own network. I was sure the script was the weak part, so I kept the exact same pixels, rewrote every word and ran it again. 93 views. 93% in-network, again. Better copy couldn't rescue a synthetic face on a cold feed, so the twin came off the feed.

Two different sneers, same lesson. People and platforms both decide in about a second whether a human showed up, and the buyer's sneer fires at effort they can tell wasn't spent on them, whatever tool was involved.

How do I decide when to use AI and when to write it myself?

I match the effort I'm given. Reciprocity. An agent reads my feed and flags what's worth my eyes, and whether it looks human-written or machine-made. Pure AI with zero thought behind it gets an AI reply. Something you clearly wrote yourself gets a real reply, written by me. Somewhere in between gets one I drafted with AI and actually read before it went out.

Openers are the fastest tell. If something starts with "I have to be honest" or promises to say the quiet part out loud, it gets the instant AI reply, because if you won't work for your impressions, neither will I.

Sounds cold. In practice it's how I protect the human part. I use AI so the human moments survive. It clears the junk, which means the people who show up as people get a person back.

I also keep score in public, losses included. We ran a live outbound test: same list, same sequence, one batch got a behavioral read of the buyer before the first email and one didn't. I predicted the read would double replies. It got zero. 0 of 24 with the read, 0 of 25 without. Opens were 71% on the read arm and 76% on the control, with bounces at 0%, so people saw it, opened it and passed. That's the message and the offer missing. With both arms at zero, the read hasn't been beaten yet, only left untested. I published the zero anyway, because a founder who only shows you the weeks that worked is running the same trick as an AI email with your first name pasted on top.

How do you use AI for sales content without sounding like AI?

Do the reading yourself and let the tool handle the typing. Concretely, starting Monday:

1. Read the human before you generate anything. Write down one thing that is only true of this person. If you can't, you aren't ready to write to them.
2. Let AI draft, then read it as the buyer. Cut any opener that could start anyone's email. "I have to be honest" and "read that again" go first.
3. Decide where your human attention is earned. Tier your messages the way I tier replies: fully human, AI-drafted and human-read, or automated. Be honest with yourself about which bucket each one is in.
4. Spend the time AI saves on the deal. The draft was never where deals died. Put the hour into the stakeholder who went quiet.
5. Keep score, flops included. If you only count the wins, you'll never learn which part was the read and which part was luck.

The read is the part we built GTM Heroes around (the relationship lens). The tool isn't the monster. The monster is the quiet question of whether you still matter, and you do, just for a different part of the job than you thought. Stop selling to the buying committee. Start selling to the humans in it.

(And yes, this is original thought and creative supported by AI.)