n8n routes customer enquiries through Telegram
Toma built an n8n customer-support workflow after enquiries arriving by email and web forms began losing context. He wanted AI assistance without handing over full control. His system captures each enquiry, assigns it a unique ID, creates a ClickUp task and sends him a Telegram notification.
x.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Complete this task from my stated requirements, keep a concise log of decisions and sources, and pause before sending messages or spending money.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- Toma (@Toma_web3)
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Toma says the workflow keeps context in one place and makes replies faster. No measured response-time or customer outcome data is published.
- Human involvement
- Toma remains involved by responding in Telegram. The post does not describe an explicit draft approval gate before customer emails are sent.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- The workflow captures enquiries from email or forms, assigns a unique ID and creates a ClickUp task.
- It sends Toma an immediate Telegram notification with the enquiry context.
- Toma replies in Telegram; the AI interprets whether to email the customer, update status or close the task.
Toma built an n8n customer-support workflow after enquiries arriving by email and web forms began losing context. He wanted AI assistance without handing over full control. His system captures each enquiry, assigns it a unique ID, creates a ClickUp task and sends him a Telegram notification.
Toma then replies naturally from Telegram. The AI reads his message and decides whether to send an email, update the enquiry status or close the thread. The workflow connects n8n, OpenAI, Telegram, ClickUp, Gmail and Google Sheets, so he can handle the conversation and its task record from one place instead of switching between tools.
Toma describes the result as faster replies and fewer lost details, but does not provide a time measurement or customer-response statistics. The human checkpoint is central: he writes the response in Telegram, and the system interprets that instruction before taking the next step. Anyone adapting this should test how it distinguishes a draft from permission to send, and log every status change so a misrouted enquiry can be recovered.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: X · Toma (@Toma_web3)
Frequently asked questions
What problem was Toma solving?
Customer enquiries arrived from email and forms, and their context was getting lost across tools.
How does the workflow use Telegram?
It sends Toma an alert with the enquiry, and he replies in Telegram; the AI can then send an email, update the task or close the thread.
Does it send replies without human input?
Toma says he replies from Telegram and the AI interprets that response. The post does not describe a separate approval step before sending, so that behavior should be tested carefully.
More details
About this account
n8n and OpenAI route enquiries, create task records and interpret Toma’s Telegram instructions to update or reply.
Reviewed the archived direct X page capture (HTTP 200, 2026-10-15) for original text, author and post date; a fresh fetch from the current web reader returned 403. Toma shares a workflow description rather than a verbatim prompt; the prompt is editorially suggested.
Credit: Toma (@Toma_web3) — X
Source published
Added to Agents In Use
Source checked
Tools and topics
n8n · OpenAI · Telegram · ClickUp · Gmail · Google Sheets
customer support · Telegram · n8n workflow
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.