If your business manages multiple websites, brands, or audiences, using separate AI agents for each one helps deliver more accurate and relevant customer support. Each AI agent can have its own dedicated knowledge sources, custom skills, and behavior tailored to a specific website or use case.
In Text, you can create and manage multiple AI agents from Team → AI agents. By combining dedicated knowledge sources, Custom Skills, and Chat routing rules, you can ensure that every visitor is connected to the right AI agent for their needs.
Multiple AI agents are separate AI assistants configured for different websites, audiences, or workflows in Text. Each AI agent can have its own knowledge sources, Custom Skills, and chat routing setup, helping it provide more accurate and context-aware support.
Using multiple AI agents helps you:
Each license includes at least one AI agent. Multiple AI agents are available on Growth and Enterprise plans.
For example, you can:
Create separate AI agents for different websites or product lines
Assign unique knowledge sources containing only relevant content
Customize skills and responses for each audience
Automatically route chats based on website domains or URLs
Using AI agents this way helps maintain consistent, context-aware conversations and improves the customer experience across all your websites.
AI agent assignment depends on your team setup and Chat routing rules.
When you create a new AI agent:
It is automatically assigned to the Chatting team
It stays in the Chatting team unless the AI agent is deleted
You can assign AI agents to different teams in the Teams section
Chat routing rules determine which team receives incoming conversations
Once a conversation is routed to a team, the available AI agent within that team can handle it
Only one AI agent can handle a conversation at a time. When multiple AI agents are added to the same team, the selected Priority determines which AI agent handles incoming chats.
Primary agents always receive chats first. Backup agents receive chats only when all Primary agents are at full capacity.
Each AI agent can handle up to 500 concurrent chats, meaning it can manage 500 active conversations simultaneously.
You can create and manage AI agents from Team → AI agents in Text.
Step 1: Create a new AI agent
Open Team → AI agents and create a separate AI agent for each website, brand, or workflow.
For example, you might create:
A sales AI agent for your ecommerce website
A support AI agent for your Help Center
A multilingual AI agent for regional websites
A lead generation AI agent that collects contact details
Step 2: Configure the AI agent
Each AI agent can have its own:
Name and avatar
Tone of voice
Custom instructions and behavior
Response length
Knowledge sources
Custom Skills
Assigned team
Supported languages
Keeping each AI agent focused on a specific audience improves answer quality and reduces irrelevant responses.
A dedicated knowledge source ensures the AI agent only answers questions related to its assigned website or audience.
Using separate knowledge sources helps:
Improve AI answer accuracy
Prevent unrelated responses
Keep product information organized
Create more consistent customer experiences
For best results:
Add only content relevant to that website
Remove outdated documentation
Keep articles organized by topic or product
Enable auto-update to keep knowledge sources accurate when your website content changes frequently
Step 3: Assign chat routing rules and start accepting chats
After configuring the AI agent, set up Chat routing rules to direct incoming conversations to the appropriate AI assistant.
Custom Skills allow you to tailor how each AI agent behaves during conversations.
You can manage skills in:
Team → AI agents → Skills
What are Custom Skills?
Custom Skills are automated workflows for AI agents that trigger when the AI agent detects a specific customer intent, keyword, context, page, location, or time-based condition. They let the AI agent go beyond replying by taking actions such as collecting details, sending messages, routing chats, tagging conversations, updating customers, or creating tickets.
In short, they help teams automate support and sales processes in plain language, reduce manual work, and deliver faster, more consistent customer experiences.
Examples of Custom Skills include:
Product recommendation guidance
Troubleshooting workflows
Lead qualification instructions
Escalation handling
Language-specific communication
Create Custom Skills for different websites
Different websites often require different communication styles and workflows.
For example:
An ecommerce website may need product recommendations and sales-focused skills
A Help Center may need troubleshooting and support workflows
A SaaS dashboard may require onboarding and account guidance
A regional website may need multilingual communication support
Using separate Custom Skills for different AI agents helps create a more personalized customer experience across websites.
After creating your AI agents, use Chat routing rules to automatically assign conversations to the correct one.
Chat routing rules automatically assign incoming chats to a specific team based on conditions such as the page URL, website domain, or customer location (country). This helps ensure visitors always interact with the most relevant AI agent.
Set up Chat routing rules
Step 1: Open routing settings
Go to Settings → Routing → Chat routing in Text.
Step 2: Create routing conditions
Set conditions based on:
Page URL
Website domain
Customer location (country)
For example:
Chats from a support domain can route to a support AI agent
Chats from an online store can route to a sales AI agent
Step 3: Choose the team
Choose the team that should receive chats matching the routing conditions. The AI agent assigned to that team will handle those chats.
Step 4: Test the routing setup
Open each website and confirm conversations are routed to the correct AI agent.
Troubleshooting AI agent routing issues
Chats are routed to the wrong AI agent
Review your Chat routing rules and verify:
Domains are configured correctly
URL conditions match the correct pages
Rule priority is set properly
AI responses contain unrelated information
Check the assigned knowledge source and remove content unrelated to that website or workflow.
AI responses sound inconsistent
Review your Custom Skills and Custom instructions to ensure the tone and behavior are aligned across conversations.
For the best AI response quality, create a separate AI agent for each website or audience, keep knowledge sources up to date, and regularly review chat routing rules and conversations after making configuration changes.
