2 min read

Making Telegram the front door for your AI agent

Making Telegram the front door for your AI agent

Hook

I get a ping on Telegram. It is not a person. It is the AI agent I deployed last week. It tells me that our lead database just grew by 12 contacts and that one of them matches a high-intent sales trigger. I reply with a short command, and the agent schedules a follow-up email. No dashboards. No email threads. Just Telegram.

Background

At 1it.pro we run a handful of internal workflows around sales ops and customer support. We used to glue together a bunch of scripts, cron jobs, and shared spreadsheets. It worked, but it was noisy.

Every time we added a new script, we also added a new notification channel. Email, Slack, a dedicated web dashboard. The team started ignoring most of them. The signal was buried under the noise.

Problem

The real problem was not automation. It was attention. We had the data, but we did not have a single place where the team could interact with the agent naturally.

Each tool had its own login. Each script had its own output format. The AI agent could generate insights, but nobody checked the web UI every hour. The agent was smart, but it was stuck inside a box that nobody opened.

Why it matters

When the agent cannot reach the person who needs the information, the whole investment in AI becomes a background process nobody uses. The cost of doing nothing is a growing pile of missed opportunities and delayed responses.

For a small team, every minute spent switching contexts is a minute lost. If the agent messages you where you already talk to your colleagues, that friction disappears.

Solution

I set up AgentForge on a small VPS with Docker Compose. The installation took about 15 minutes. Then I opened the web UI to configure the agent’s Telegram integration.

The setup was straightforward: create a Telegram bot via BotFather, copy the token, paste it into the web UI. After that I defined a few commands the agent would recognise in Telegram chats.

  • 📞 /leads — fetch new leads from the CRM and list them
  • ⚠️ /alerts — show any recent system anomalies
  • 📅 /remind — create a task for a specific time

The agent runs inside a Docker container, and the web UI allows me to edit its behaviour without touching YAML files. I added a short prompt that tells the agent to respond only in plain text and to confirm actions before executing them.

Telegram became the single interface. No new dashboard. No extra tab. Just the chat app the team already had open all day.

Result

Within a day, the agent was handling basic support triage and lead notifications directly in our team chat. People started replying to it. They asked it to reschedule reminders. They asked it to pull up customer notes.

The key metric for me was this: the number of manual follow-ups dropped by about 40% in the first week. We were not monitoring more — we were replying faster. The agent was no longer a side project. It was part of the daily flow.

Telegram is not just a notification channel anymore. It is the front door for the agent, and the agent does real work.

AgentForge

AI agents for real work. If you want to see how this works in practice, visit our product site and blog.

Share
Serhii

Serhii