A post by Max on IndieHackers went viral this week with an uncomfortable data point: four months spending $400 a month on six AI tools to automate his SaaS distribution, and zero paid leads attributable to that investment. After an honest audit of 40 weekly tasks, he trimmed the stack to $19 a month and doubled output. The difference wasn't better tools, it was separating well what to automate and what never to. This article applies that same logic to blog distribution in 2026.
The problem with "automated distribution" as it's sold
The market for AI creator tools promises three things: save time, maintain consistency, and scale channels. All three are true for part of the work. The problem is that most creators accept the full package without auditing which tasks tolerate automation. The result: generic outreach emails, empty LinkedIn comments, follow-ups that lose context, and a feeling of "I'm scaling" that doesn't translate into subscribers.
The audit any creator should run this Monday
List the 40 distribution actions you do in a typical week. For each, tag one of three categories:
| Type | Examples | Automate? |
|---|---|---|
| Mechanical-repetitive | Crossposting to social, OG image generation, video cuts, newsletter reminders | Yes, no thought needed |
| Drafting-reviewable | First draft of a post, Twitter thread skeleton, email subject lines | Yes, with human editing at the end |
| Voice/relationship-dependent | Responding to a reader's follow, commenting on a peer's post, follow-up to an interview, collaboration pitch | Never |
The operating rule: if the task requires remembering something specific about the person on the other side, or adjusting tone to context, it's out of the "automate" list.
The real stack (that fits in $20 a month)
What to automate
- Your blog platform's native publishing scheduler (don't buy a separate Buffer if your CMS already has it)
- Automatic OG image generator (Vlogerly does it by default; on other platforms, Vercel OG or similar)
- Newsletter send reminders (a simple cron is enough, you don't need a SaaS)
- Initial drafts with an LLM (use them as a scratch, not as output)
What NOT to automate
- Replies to comments on your blog or social
- DMs to specific people who read you
- Pitches to journalists or bigger newsletters
- Follow-ups after any 1:1 conversation
The most painful lesson from the original post: an auto-send script didn't respond for 5 days to a journalist who had shown interest. By the time the founder noticed, the press opportunity was dead. Automation has invisible opportunity cost.
The "never-write list": the voice filter
If you're going to use an LLM for drafts, give it an explicit never-write list. It's a 1-page document with phrases and patterns the AI must avoid because they're not your voice. You write it once and put it in the system prompt every time. Typical examples for a blog:
- Don't start paragraphs with "In today's fast-paced world"
- Don't use "In conclusion" or "To wrap up"
- No emojis in body text
- No em-dash "—" (reads as AI to many readers)
- No clickbait promises in headlines ("This will change your blog forever")
Every creator should have their own. Takes ten minutes to write and drastically lowers the ratio of unusable drafts.
How to measure if your stack is working
Vanity metrics (impressions, reach) are noise. The signal metrics are three:
- New newsletter subscribers per week with clean attribution to the channel (utm on every distribution link)
- Human replies initiated by readers after reading a post (a comment, an email, a DM). Few but real
- Average time spent editing vs raw production. If you add tools and editing drops, you're heading toward generic content
"Audit your 40 weekly distribution actions before buying another AI tool. Separate tasks into 'AI can handle' and 'AI cannot without killing the channel.' The second category is the one that converts" (summary of the original IndieHackers post).
Conclusion
The expensive mistake isn't buying tools, it's delegating the wrong tasks. For a blog in 2026, repetitive mechanics do scale with AI; relationship with your audience does not. Before adding the next SaaS to the stack, ask yourself which of your tasks require memory of human context. Those are the ones that generate subscribers who stick.
If your blog platform already includes scheduling, automatic OG images, and built-in newsletter, you can keep your distribution stack cheap and dedicate the difference to human work. Try Vlogerly for free and remove five SaaS from your stack at once.
