Skip to main content

Agent Behavior and Response

Agents in Telex are designed to behave like intelligent digital teammates. Their responses are shaped by a combination of configuration inputs and contextual awareness, allowing them to interact naturally and execute tasks with purpose.

What Shapes Agent Behavior

Each agent’s behavior is determined by:

  • System Prompt
    The foundational instruction that defines the agent’s personality, tone, and role. This prompt guides how the agent interprets messages and frames its responses.

  • Role Configuration
    Roles influence the agent’s domain expertise and conversational style. For example, a support agent may respond with empathy and troubleshooting steps, while a lead generation agent may ask qualifying questions.

  • Interaction Context
    Agents behave differently depending on where they’re used:

    • In channels, they respond publicly and contribute to group workflows.
    • In direct messages, they operate more privately and can focus on individual tasks.
  • Task Lists (coming soon)
    These will allow agents to follow structured workflows, enabling multi-step execution and conditional logic.

  • Skills (coming soon)
    Modular capabilities that define what an agent can do — such as sending emails, summarizing threads, or generating reports.

Response Style

Agents respond using natural language and aim to be:

  • Relevant — staying on topic and aligned with the prompt
  • Helpful — offering actionable insights or completing tasks
  • Context-aware — adapting to the conversation flow

Agents do not require special commands. Just speak naturally, and they’ll respond based on their configuration.