Connecting Telegram to AI: every employee their own bot
Telegram can be connected to IntraGPT as a channel. Each employee creates their own bot with Telegram, pastes the token into the dashboard and from then on chats with their own assistant. An incoming message goes to that personal conversation and the answer comes back in the same chat, with the knowledge and the rights belonging to that employee.
What you do with the Telegram connection
A personal bot per employee
You create your own bot with Telegram and paste the token into IntraGPT. From then on you talk to your assistant, not to a shared company bot in which conversations run into each other.
Questions and answers in the same chat
What you send to the bot reaches the assistant and the answer appears in the same conversation. Follow-up questions belong to the same thread, so the context stays put.
Visibly working
While the answer is being prepared the bot shows it is typing. That way you know your message arrived, even when the answer takes a moment.
Seeing whether the connection holds
The dashboard shows whether the connection is active and lets you refresh its status. When something does not work you do not have to guess whether it is the bot or the message.
Setting a workflow in motion
An incoming message can do more than produce an answer: it can start a defined workflow, so a report ends up in the right place by itself.
What employees use it for
Engineering
- Situation
- An employee works across changing devices and wants the same assistant at hand on all three.
- What the agent does
- He connects his own bot and uses it on his phone, tablet and laptop, with the same conversation history.
- Result
- One assistant that travels everywhere, without logging in again per device.
Project organisation
- Situation
- Small questions come up during work that are not big enough to call a colleague about.
- What the agent does
- The question goes to the personal bot, which takes the answer from the connected sources.
- Result
- Answers at the moment the question arises, instead of a list that evaporates by the end of the day.
Field service
- Situation
- A report from the road has to land in the right system, but nobody opens a form for it.
- What the agent does
- The message to the bot sets a workflow in motion that forwards the report and asks for the missing details.
- Result
- Reports that arrive complete, without extra effort from the person reporting.
Management
- Situation
- A director wants to check something quickly without starting up a workstation.
- What the agent does
- He sends the question to his own bot and gets the answer back with its source.
- Result
- An answer within a minute, with the same boundaries as in the web environment.
Connected in three moves
-
1
Creating your own bot
With Telegram's bot manager you create a bot and choose the name. You receive a token that belongs to you alone. You share that token with nobody.
-
2
Putting the token in the dashboard
In IntraGPT you paste the token into the connection screen. It is stored encrypted with your user, not shared centrally, and IntraGPT checks straight away whether the bot is reachable.
-
3
Deciding what the assistant may do
The assistant behind your bot has your rights and your sources. What you may not view in the web environment does not surface through Telegram either.
-
4
Disconnecting again
If you want to stop, you break the connection in the same screen. After that nothing comes in. The bot stays yours and you can remove it at Telegram itself.
Why a bot per person is safer
With a shared company bot, conversations and rights run into one another. Giving every employee their own bot takes that question off the table.
Where the data lives: Nederland of de EU
Separated per employee
Each connection belongs to one user. There is no shared pool of conversations several people can look into and no shared token doing the rounds.
The token stays shielded
The bot token is kept encrypted and cannot be read back after entry. If you lose it, you create a new one with Telegram and replace it.
The same rights as in the chat
The assistant knows no separate set of rights for Telegram. What an employee may see in the web environment, they see here, and nothing beyond it.
The thinking happens with us
Answers are composed on our own server in the Netherlands. The content does not travel to an American model vendor and is not used for training.
Questions about the Telegram connection
Does every employee have to create their own bot?
Yes, and that is on purpose. Everyone keeps their own conversations and their own rights, and a connection can be undone per person. Creating one takes a couple of minutes and goes through Telegram's own bot manager.
Does IntraGPT work with Telegram?
Yes. Telegram can be connected as a channel: you paste your own bot token into the dashboard and then chat with your assistant in Telegram, with the same knowledge as in the web environment.
Can I connect Telegram to ChatGPT?
There is no built-in connection between Telegram and ChatGPT, and sending company information to a public service is not what you want. With IntraGPT the bot hangs off your account in your own environment, with your sources and your boundaries.
What is the difference with the WhatsApp channel?
With WhatsApp you pair an existing phone number using a QR code, and that number has to stay online. With Telegram you work with a bot that exists independently of a phone, one per employee. Telegram is therefore steadier for personal use, WhatsApp is handier when you specifically want to use a familiar company number.
Can I disconnect my bot again?
Yes, in the same screen where you connected it. After that nothing reaches the assistant. The bot itself keeps existing at Telegram, so you can reconnect later or remove it there.
Does it work in a group as well?
A bot does not see everything said in a Telegram group by default; that is a Telegram setting. For work a personal conversation with your own bot tends to be more pleasant anyway, because the answer then reaches only the person who asked.
What does a Telegram connection cost?
Connecting is part of the track we go through together. What drives the work is which assistant sits behind it, which sources it may use and whether workflows hang off it. Thirty minutes is enough to make that clear.
How long does setup take?
Creating a bot and pasting the token takes a couple of minutes per employee. The real work sits in the assistant behind it: which knowledge it may use and what it may do on its own.