I Set Up AI Automations to Get My Time Back. Then I Noticed the Slop.
Three months ago I automated most of my content pipeline. Emails out. Social posts scheduled. Blog drafts in the queue. Inboxes triaged. I woke up one morning and had two hours I didn't know what to do with.
That was the good part.
The bad part started about four weeks later, when I went back through everything I'd published and noticed something uncomfortable: it all sounded the same. Not just similar — the same. The same rhythm. The same transitions. The same five phrases I'd apparently taught the AI to recycle.
The automation worked. The output was technically correct. The problem was that "technically correct" is a low bar when you're trying to build something people actually want to read.
What "getting time back" actually looks like
I want to be specific here, because "AI automations" means a hundred different things depending on who's selling it.
My setup:
- Content drafts: A script reads my voice notes and spits out a rough first paragraph. I rewrite most of it.
- Social scheduling: Buffer queue populated from a template that pulls the week's theme. I still write each caption — the automation just removes the "when do I post this" decision.
- Email triage: A classification script flags what's urgent and what's noise. Still my hands on every response.
- CRM updates: Call notes go in, fields update automatically. This one actually works perfectly.
The net result was roughly 90 minutes a day. I used them badly at first — doom-scrolled, then felt guilty — then slowly started using them for actual deep work.
Where the slop crept in
The first sign was the LinkedIn posts. They were getting engagement (somewhat). They were getting reach (sort of). They were saying something (occasionally).
They were not saying anything I would have said.
There's a specific texture to content written by an AI that hasn't been edited by a human who cares. It's not wrong, exactly. It's not even vague. It's just… flattened. Every edge rounded off. Every opinion that might land wrong removed preemptively.
The automation was producing content that was acceptable. Acceptable is the enemy of interesting when you're building an audience of people who have thousands of other things competing for their attention.
What I changed
I started treating AI output as a rough draft, not a finished product.
That sounds obvious. It wasn't obvious to me at 11pm when the automation had knocked out six social posts and I was tired and it was fine and I just hit schedule.
The rules I set for myself:
- Every AI-generated sentence gets one manual edit. Doesn't have to be big — a word change, a different rhythm, my actual voice reasserting itself. Just enough to make it mine.
- The hook is always mine. The opening line of anything that goes out under my name is written by me, first. The AI fills in the body, not the beginning.
- I fact-check the specifics. Numbers, names, product versions — the things AI hallucinates most reliably. I caught two product names wrong in one month. Both were embarrassing.
- I read it out loud before posting. This catches the flatness. If it sounds like something I'd say in conversation, it goes. If it sounds like a LinkedIn post, it gets rewritten.
Getting time back from automation is real. Using that time to publish more noise defeats the point.
The actual win
After I stopped treating AI output as publishable and started treating it as raw material, something shifted. The content was still faster to produce. But it started sounding like me again.
My engagement didn't spike. It didn't crash either. But the messages changed — people started responding to specific things I'd said, not just reacting to a pattern.
The time I'd recovered was still there. But I was now using it to think, which is the thing AI can't do for you.
The bottom line
AI automations can give you time back. They gave me back two hours a day, which over three months is roughly one full work week of attention I didn't have before.
Whether that time becomes something valuable depends entirely on what you do with it. If you use it to produce more content faster, you'll have more content. If you use it to actually think about what you're building, you might end up with something worth reading.
The automation is a tool. The judgment about what to do with the time it frees up? That's still yours.