Pular para o conteúdo
SynapsAI Docs

Agents

Este conteúdo não está disponível em sua língua ainda.

Agents are the core building blocks of SynapsAI. Each agent is an isolated workspace with its own knowledge base, behavior configuration, extensions, and conversation history. You can run multiple agents under a single account — one per brand, department, or use case.

What is an agent?

Section titled “What is an agent?”

An agent is an AI-powered assistant that:

  • Answers questions using a knowledge base you provide (RAG).
  • Follows rules defined in a system prompt — tone, boundaries, fallback behavior.
  • Connects to channels like your website (Live Chat), WhatsApp, or the API.
  • Uses extensions to perform actions like booking appointments or looking up products.
  • Hands off to humans when it can’t resolve a question.

Each agent is completely isolated. Documents uploaded to one agent are not visible to another. Conversations, settings, and extensions are all agent-specific.

Agent lifecycle

Section titled “Agent lifecycle”
  1. Create — give it a name and language. See Create an Agent.
  2. Add knowledge — upload documents, FAQs, and website content. See Knowledge Base.
  3. Configure behavior — write a system prompt and choose a model. See Configure Behavior.
  4. Test — use the Playground to verify response quality.
  5. Deploy — connect channels like Live Chat or WhatsApp.
  6. Monitor — review Conversations and dashboard metrics.
  7. Iterate — refine knowledge, adjust the prompt, add extensions.

Managing multiple agents

Section titled “Managing multiple agents”

If you need separate agents for different purposes, you can create as many as your plan allows. Common setups:

ScenarioAgents
Single product, one audience1 agent
Multiple brands under one company1 agent per brand
Support + Sales separation1 agent per function
Agency managing multiple clients1 agent per client

Switch between agents using the Agents list in the sidebar. Each agent has its own dashboard, knowledge base, and settings.

Agent-level sections

Section titled “Agent-level sections”

Once you select an agent, the sidebar shows these sections:

  • Dashboard — metrics and activity overview.
  • Knowledge Base — documents and sources that power RAG.
  • FAQ — structured question-and-answer pairs.
  • Playground — test your agent in real time.
  • File Inventory — manage uploaded files.
  • Catalog — product and service items.
  • Conversations — review past and active chats.
  • Triggers — automation rules.
  • Extensions — third-party integrations.
  • Operators — team members and roles.
  • Agent Settings — model, prompt, and configuration.

Next steps

Section titled “Next steps”