Marketing

48 of 48 Review Replies Opened 'Thank You For': Why a Better Prompt Wasn't Enough for His n8n Review Bot

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

What they did

He ran 8 realistic Google reviews through an n8n review-reply workflow 6 times each (48 drafts). Every draft opened with 'Thank you for', 31 used stock phrases and 20 didn't mention anything the reviewer said. A stricter prompt helped only partly (29 of 48 still opened 'Thank you for') and made new problems: 3 mentioned refunds or offers and 6 left a placeholder or a made-up phone number. What worked was a code check that bans stock phrases, money promises and placeholders, requires a word from the review, adds the sign-off itself, and always sends 1-2 star or health and legal reviews to a human.

How it went

dev.to post by crzy_undefinedundefined (Oct 26, 2026): 48 drafts per version; 48/48 'Thank you for' openers, the prompt fix cut it to 29/48 but added 6 placeholder or invented-number drafts; a code gate decides what a human checks. Tested on a small local model and a made-up cafe. Always give the real contact details so it can't make one up

Worth knowing

  • dev.to post by crzy_undefinedundefined (Oct 26, 2026): 48 drafts per version; 48/48 'Thank you for' openers, the prompt fix cut it to 29/48 but added 6 placeholder or invented-number drafts; a code gate decides what a human checks.
  • Tested on a small local model and a made-up cafe. Always give the real contact details so it can't make one up.

Prompt

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

Draft replies to my new Google reviews as [owner name] of [business], under 60 words. Mention one specific thing the reviewer wrote, don't open with 'Thank you for', and never offer refunds, discounts or admit fault. Use only this contact method: [real phone/email]. Put every 1-2 star review, and any mention of illness, safety, legal threats or allergies, in a 'needs me' list instead of drafting a public reply.

Recommended

Try it in Grok Bot

If Grok Bot answers your reviews, give it your real contact details and have it hold 1-2 star replies for you.

  1. Copy the prompt (button below).
  2. Open Grok Bot chat and paste it. Replace any [bracketed] placeholders.
  3. Connect the listed tools (Email) 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

Tools: Email · 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 by crzy_undefinedundefined (Oct 26, 2026): 48 drafts per version; 48/48 'Thank you for' openers, the prompt fix cut it to 29/48 but added 6 placeholder or invented-number drafts; a code gate decides what a human checks. Tested on a small local model and a made-up cafe. Always give the real contact details so it can't make one up

Source: Web · 2026-10-26 · 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