I'm Pavlo, an n8n automation freelancer. Below are eight workflows that run in production on my own server right now. They handle orders, email funnels, bookings, alerts and content publishing without anyone clicking a button.
Problem: orders arrive, stock lives in a spreadsheet, someone has to check both.
Workflow: a webhook receives the order, validates it and checks stock in Google Sheets. In stock, it logs the order, updates inventory, emails the customer and pings Telegram. Out of stock, it sends a different email and a Telegram alert.
Problem: collect emails legally and follow up without a paid email tool.
Workflow: a signed confirmation link (HMAC) verifies the address, then sends the free sample. A daily job sends the follow-up sequence to confirmed contacts, and every mail carries a signed one-click unsubscribe.
Problem: small service businesses lose clients who message outside working hours.
Workflow: customers chat in plain language and an LLM works out what they want. The workflow finds free slots, proposes them, books the slot and prevents double-booking. The owner manages hours and blocked time by chatting too, protected by a PIN.
Problem: workflows fail silently and you find out days later.
Workflow: one Error Trigger workflow catches failures from every other workflow. It logs them to a sheet, sends a Telegram message, and emails you as well when the failure is critical.
Problem: comments pile up and replying by hand costs hours.
Workflow: every 30 minutes during the day it fetches new replies to your posts and answers the relevant ones after a random delay. It checks the API token first and warns you on Telegram before it expires. Every action is logged to chat.
Problem: a buyer pays for a custom invitation and waits for you to make it.
Workflow: a Gumroad sale webhook extracts the buyer's invitation code, renders the PDF, emails it through Resend, notifies you on Telegram and saves the customer to a sheet.
Problem: a daily news video takes hours of research, voice-over and editing.
Workflow: it reads RSS feeds, removes duplicate stories, has an LLM write the script, then generates the voice-over, images and subtitles. ffmpeg renders the video and a short, and the result goes into a publishing queue.
Problem: publishing an article means uploading files and updating the sitemap.
Workflow: it validates the incoming article, writes the HTML file to the web server over SSH, updates the sitemap and pings Google. One call and the page is live.