Introduction
Chat agents provide the same intelligent automation as voice agents, with one key difference: they operate across text-based channels.Identical Layout to Voice
Chat agents share the same configuration interface as voice agents, making it easy to switch between the two or run both simultaneously.Platform Integrations
The key extra feature – Plug straight into multiple platforms for truly omnichannel conversations:- Instagram - Respond to DMs automatically
- Facebook Messenger - Handle messages 24/7
- SMS - Text with prospects
- WhatsApp - Global messaging support
- Web-chat - Widget for your website
- Email - Automated email responses
Model Options
Choose from several AI models to power your chat agent:- GPT-4 - Current favorite for most use cases
- GPT-4o - Enhanced version with improved performance
- GPT-4.1 - Recommended balance of speed, cost and intelligence
- GPT-3.5 - Faster and more economical for simple tasks
- Claude - Alternative model for specific needs
Configuration Options
Triggers for Campaigns & Keyword Flows
Add a trigger word or phrase (for example “start”).The agent only wakes up when the incoming message matches the trigger. Perfect for:
- “DM me START to claim your voucher” Instagram reels
- Keeping the AI silent while you chat to clients manually
- Campaign-specific automated responses
Delay Settings
Minimum delay and maximum delay add a random pause before each reply. Use 30–120 seconds if you want the conversation to feel human rather than instant.Knowledge Base
Choose one of three methods to give your agent information:- File upload – Drop PDFs, Word docs, CSVs, plain-text files—anything
- Website scrape – Paste a URL; Callgency AI scrapes the site (via Firescraper) into tidy Markdown
- Plain text – Paste raw text directly
Testing & Sharing
Test Agent – Chat straight inside the dashboard to iterate quickly. External Preview (arrow-box icon) – Generates a unique link you can send to a client for live trials. All conversations and tool calls land in Transcripts, so you can review inputs, outputs, and tweak with real data.Next Steps
- Learn about Prompt Engineering to create effective chat agents
- See Chat Agent Examples for real-world implementations
- Explore GoHighLevel Integration for advanced automation

