Skip to main content
Every ASCN Agent is three layers. Understanding them means understanding how to configure an agent for any task. In the side menu you see three key sections: Brain (①), Channels (②) and Tasks (③) — each owns its layer. Agent structure in the side menu

Brain — the control center

Brain defines the agent’s personality, knowledge and abilities. Here you configure:
  • Integrations — which services the agent can access: Gmail, Google Calendar, Google Drive, Slack, GitHub
  • Skills — reusable behaviour scripts for typical tasks
  • Knowledge — who the agent is, how it talks, which files and documents it uses
  • Memory — what the agent remembers about you between sessions
The Brain is a one-time setup. Once you’ve connected integrations and filled in knowledge, the agent is ready to work without further instructions.

Tasks — automations

Tasks are what the agent does on its own, without you. Two kinds:
TypeExample
ScheduledEvery morning at 8:00 — summary of unread email
TriggeredNew email arrives → agent prepares a draft reply
Tasks are created in chat — just describe what to do and when.

Channels — where you talk

Channels are where you talk to the agent and where it replies. Available channels:
  • ASCN chat — web interface of the platform
  • Telegram — a bot you create via BotFather
  • Slack — your team workspace
Connect Telegram so you can control the agent and get notifications right in the messenger — without opening a browser.

How the layers work together

Example: a morning email agent
  1. Brain — connected to Gmail and Google Calendar, knows your work style
  2. Tasks — every day at 8:00 the agent reads your inbox and builds a summary
  3. Channels — the summary comes to you in Telegram
You wake up and already know what arrived overnight.

Integrations

Start with the Brain — connect your first integration.