Workflows
Six workflows, ready to import
Each file passed a structural check (valid nodes and connections, no embedded secrets). Credentials are referenced by name only, so you connect your own accounts.
01Website lead qualification
Every website enquiry is scored, logged and the hot ones reach your phone in seconds.
- The business problem
- Enquiries from the contact form sit in an inbox until someone reads them, and the serious ones wait as long as the weak ones.
- What it does
- A form on your site posts the enquiry to an n8n webhook.
- An LLM scores intent and fit from 1 to 10 and returns a structured answer: score, priority, reason, next step.
- The result is written as a new row in Google Sheets.
- If the lead is marked hot, a Telegram message reaches you straight away.
- The webhook answers the form so the visitor sees a confirmation.
- Tools and nodes
- WebhookLLM chain with output schemaGoogle SheetsIfTelegramRespond to Webhook
- What you need
- Credentials, selected in n8n by name: OpenAI, Google Sheets, Telegram. A sheet tab named Leads with the columns name, email, company, message, score, priority, reason, next_step, received_at; your sheet ID and Telegram chat ID.
FlowMainWebsite form→Lead webhook→Qualify lead (LLM)→Build row→Sheets append→Hot lead?
HotTelegram alert→Respond OK
Download workflow JSON ↓
02Gmail triage with reply drafts
The inbox sorts itself and every new lead already has a reply draft waiting.
- The business problem
- Sales enquiries, support questions, invoices and spam arrive in one inbox, and the person reading it spends the first hour sorting.
- What it does
- A Gmail trigger picks up each new message.
- An LLM returns a category (lead, support, invoice, spam) with a confidence value.
- A switch applies the matching Gmail label.
- For leads, a second LLM writes a reply and saves it as a draft in the same thread.
- Nothing is sent automatically; you read the draft and press send.
- Tools and nodes
- Gmail TriggerLLM chain with output schemaSwitchGmailGmail draft
- What you need
- Credentials, selected in n8n by name: OpenAI, Gmail. Four Gmail labels (lead, support, invoice, spam), selected in the label nodes.
FlowMainNew e-mail→Classify (LLM)→Route by category
LeadLabel lead→Draft reply (LLM)→Save draft
Download workflow JSON ↓
03Telegram FAQ assistant with human handoff
Customers get instant answers from your FAQ and a human steps in when they ask.
- The business problem
- The same ten questions arrive in your messenger all day, and the questions that need a person get buried among them.
- What it does
- A Telegram bot receives the customer message.
- If the text contains a handoff keyword, staff are notified and the customer is told a person will reply.
- Otherwise the workflow fetches your FAQ from a URL.
- An AI agent with per-chat memory (last 8 messages) answers strictly from that FAQ.
- The answer is sent back in the same chat.
- Tools and nodes
- Telegram TriggerIfHTTP RequestAI AgentWindow buffer memoryTelegram
- What you need
- Credentials, selected in n8n by name: OpenAI, Telegram. A bot token from BotFather, your FAQ hosted as plain text or markdown at a URL, and the staff chat ID.
FlowHandoffTelegram message→Keyword found?→Notify staff→Tell customer
AnswerNo keyword→Fetch FAQ→Prepare context→FAQ agent + memory→Send answer
Download workflow JSON ↓
04Booking with availability check
Booking requests are checked against your availability sheet and confirmed automatically.
- The business problem
- Customers ask for a time slot, someone checks the calendar by hand, and double bookings happen when two requests overlap.
- What it does
- A booking request arrives at a webhook with name, e-mail, slot and service.
- The workflow reads your availability sheet and checks the requested slot.
- If the slot is available it is marked booked in the sheet.
- The customer gets a confirmation e-mail and you get a Telegram alert.
- If the slot is taken, the webhook answers with up to three alternative slots.
- Tools and nodes
- WebhookGoogle SheetsCodeIfGmailTelegramRespond to Webhook
- What you need
- Credentials, selected in n8n by name: Google Sheets, Gmail, Telegram. A sheet tab named Availability with the columns slot, status and booked_by, one row per bookable slot.
FlowAvailableBooking webhook→Read sheet→Check slot→Available?→Mark booked→E-mail customer→Telegram alert→Respond confirmed
TakenSlot taken→Respond 409 + alternatives
Download workflow JSON ↓
05Daily competitor and news digest
A short morning digest of what competitors and your industry did, without opening ten tabs.
- The business problem
- Keeping an eye on competitors and industry news means a daily round of tabs that is skipped as soon as the week gets busy.
- What it does
- A schedule trigger runs every morning at 08:00 in the workflow timezone.
- It reads a news RSS feed and fetches a competitor page in parallel.
- The two sources are merged into one input.
- An LLM writes a five-bullet digest with one suggested action.
- The digest is delivered to Telegram.
- Tools and nodes
- Schedule TriggerRSS ReadHTTP RequestMergeCodeLLM chainTelegram
- What you need
- Credentials, selected in n8n by name: OpenAI, Telegram. Your own RSS feed and competitor URLs, and the Telegram chat ID.
FlowSourcesEvery morning→Read news RSS→Fetch competitor page
DigestCombine sources→Build input→Summarise (LLM)→Send to Telegram
Download workflow JSON ↓
06Customer review analyzer
Every review is tagged by sentiment and topic, and you get a weekly report on what customers think.
- The business problem
- Reviews pile up in a spreadsheet or inbox, and patterns only become visible when someone has time to read them all.
- What it does
- A Google Sheets trigger fires on each new review row.
- An LLM returns sentiment, topic, urgency and a one-line summary.
- The analysis is stored in a second tab.
- Urgent negative reviews trigger an immediate Telegram alert.
- Every Monday a second branch aggregates the last 7 days and an LLM writes the weekly summary, sent to Telegram.
- Tools and nodes
- Google Sheets TriggerGoogle SheetsLLM chain with output schemaIfSchedule TriggerCodeTelegram
- What you need
- Credentials, selected in n8n by name: OpenAI, Google Sheets, Telegram. Two sheet tabs: Reviews (review, author, date) and Analysis (review, author, sentiment, topic, urgency, summary, analysed_at).
FlowEach reviewNew review row→Analyse (LLM)→Build row→Save analysis→Urgent negative?→Telegram alert
WeeklyEvery Monday→Read analysis→Aggregate 7 days→Write summary (LLM)→Send summary
Download workflow JSON ↓