Marketing

118 of 120 AI Social Posts Were Bad and 93 Forgot the Link: An n8n Test, and Why Code Checks Still Missed False Numbers

  • Level Intermediate
  • Setup1–3 hrs (estimated)
  • Evidence Strong evidence
  • CostNot stated in source
  • No-codeNo
n8nSource-linked

What they did

He built a blog-post-to-5-platforms social workflow in n8n two ways and ran both on 8 posts, 3 times each. The one-prompt version produced 118 bad posts out of 120, and 93 left out the blog link. The version with code-checked rules marked 77 posts ready and sent 43 to a human, and none of the 77 failed an independent checker. But when he read them, 10 of 21 in the first round got the blog post wrong, including saying churn rose from 6.1% to 4.3%. His conclusion: code can check format, not truth.

How it went

dev.to post (Oct 29, 2026): n8n test on 8 posts x 3 runs; one-prompt version 118/120 bad (93 missing the link); validated version 77 ready, but 10 of 21 reviewed misstated the post. He used a small local model on purpose, so a bigger model may fail less. A human still has to check meaning and numbers

Worth knowing

  • dev.to post (Oct 29, 2026): n8n test on 8 posts x 3 runs; one-prompt version 118/120 bad (93 missing the link); validated version 77 ready, but 10 of 21 reviewed misstated the post.
  • He used a small local model on purpose, so a bigger model may fail less. A human still has to check meaning and numbers.

Prompt

Cleaned version rewritten from the source — not a direct quote.

Turn this blog post [title, URL, text] into posts for [X, LinkedIn, Threads, Bluesky]. Each post must include the exact URL, stay within that platform's length and hashtag limits, open differently from the others, and only use numbers that appear in the post. After writing, check every post against these rules and list any that fail. I'll review all of them before scheduling.

Recommended

Try it in Grok Bot

Grok Bot can do this too: ask it to write the posts and then check each against the same rules before you approve.

  1. Copy the prompt (button below).
  2. Open Grok Bot chat and paste it. Replace any [bracketed] placeholders.
  3. Connect the listed tools (as needed) as connectors when Grok Bot asks.
  4. If the task repeats, ask: "Make this a daily routine."

Alternatives: the same prompt works in n8n, Claude, ChatGPT Agent, OpenClaw.

Set it up

Stack: n8n

  1. Read the original source for full context.
  2. Connect the tools your agent needs.
  3. Paste the prompt and replace [bracketed] placeholders.
  4. Run once manually and check the output before scheduling it.

Generic suggested steps, not from the source.

Evidence

dev.to post (Oct 29, 2026): n8n test on 8 posts x 3 runs; one-prompt version 118/120 bad (93 missing the link); validated version 77 ready, but 10 of 21 reviewed misstated the post. He used a small local model on purpose, so a bigger model may fail less. A human still has to check meaning and numbers

Source: Web · 2026-10-29 · Verified by us: not yet · Report broken or outdated

Related use cases

Inbox & communicationWeb · 2026-10

12 of 20 Prompt Injections Worked on His n8n Email Bot, Then 0 of 60: Let Code Make the Decisions, Not the Model

He built an n8n workflow where a small local model reads customer emails and decides priority, refunds and replies, then attacked it with 20 injection emails, 3 runs each. 12 of 20 worked on the plain version (35 of 60 runs), and the usual 'don't follow instructions in the email' prompt made it slightly worse (13 of 20). Taking every decision away from the model fixed it: the model only fills in a form, and a code step strips hidden HTML, keeps secrets out of the prompt and decides what happens. That got 0 of 60. The cost: only 19 of 36 normal emails were answered automatically, down from 31.

  • Level Advanced
  • Setup1–3 hrs*
  • Evidence Strong evidence
n8nSource-linked
Inbox & communicationReddit

6-class email router for a 22-person consultancy

An IMAP trigger feeds an OpenAI classifier that routes 80–120 emails a day to Slack, Pipedrive and Zendesk. It hit 94% accuracy on 500 emails and saves 11 hours a week.

  • Level Intermediate
  • Setup1–3 hrs*
  • Evidence Some evidence
n8nSource-linked
MoneyReddit

Invoice capture for a distributor: 4m12s down to 38s each

A shared Gmail label feeds an extractor, totals are validated to €0.01, low-confidence fields go to a human, and results push to Sage. 87% need no human touch.

  • Level Intermediate
  • Setup1–3 hrs*
  • Evidence Some evidence
n8nMakeSource-linkedNeeds verification