Skip to main content

AI Workforce Overview

AI Employees are digital team members you can configure to automate conversations, answer customer questions, capture leads, book appointments, and more. Each works well out of the box, but can be tailored to fit your business needs.

All AI Employees share a common structure, with configurable elements like their name, purpose, communication channels, capabilities, and knowledge sources. Specific roles (like Chat Receptionist or Voice Receptionist) may include additional options, but the basics are always the same.

See these guides for step-by-step setup or details on specific AI Employees:

For more detail on setup:

If you're just getting started, review the sections below for an overview of the main elements you'll configure for any AI Employee.

AI Employee elements​

AI Employees all share a similar structure—no matter what role they play in your business. This applies to pre-built AI Employees (eg. Chat Receptionist, Voice Receptionist) and Custom AI Employees you create. Each one comes with common parts you can review and configure, such as their name, purpose, communication channels, capabilities, and what knowledge they use to answer questions.

Profile​

The Profile section is where you set the basics for your AI Employee—how they appear to your team and how they introduce themselves to customers.

  • Name & Avatar:
    What this is: The display name and image used for your AI Employee across your channels.
    What to do: Choose a friendly, memorable name and a profile image (like your logo or a custom avatar) that fits the role or your brand. This helps customers and your team instantly recognize who they're interacting with.

  • Purpose:
    What this is: The job description and behavioral "guidelines" for your AI Employee. This sets the tone, goals, and important instructions for every conversation.
    What to do: Write a clear, short summary of what your AI Employee should do. Be specific about their tone, greeting, and key tasks.

Writing better instructions

Use bullets or numbered lists to make instructions clear and scannable. Be specific about what you want, not just general goals.

Good example:

  • "Introduce yourself as the business receptionist."
  • "Speak in a friendly, concise way."
  • "Always collect name, email, and phone before ending the conversation."
  • "If someone asks about pricing, direct them to our website pricing page."

Avoid vague instructions:

  • ❌ "Be helpful and professional."
  • ✅ "Greet customers warmly by name and ask how you can help them today."

Channels​

The Channels section lets you choose where your AI Employee will work for you.

What this is: Channels are the different places or platforms where your AI Employee can interact with your customers—like website chat, SMS, phone calls, or even Google Meet.

What to do:
Review the available channels for your AI Employee's role. Turn on only the channels that make sense for your business right now. For example, you might want your AI to respond on your website's chat widget, via text messages, or answer phone calls.

  • Website chat Widget
  • SMS (text messages)
  • Phone calls (voice)
  • WhatsApp
  • Facebook
  • Instagram

You can enable or disable channels at any time. Start with one or two channels while you get comfortable, then add more as you see results.

Channel assignment requirements

Some channels require specific products to unlock. For example, assigning a Custom AI Employee to web chat requires Conversations AI Standard, Pro, or Premium. Pre-built AI Employees may have different channel availability based on their role.

Capabilities​

The Capabilities section is where you decide what your AI Employee can actually do for your business.

What this is:
Capabilities are like "skills" or "tasks" you can turn on or off, depending on the role of your AI Employee. Each capability tells your AI which actions it should take or what outcomes to prioritize—such as capturing leads, booking appointments, answering FAQs, or performing a role-specific function.

What to do:
Review the available capabilities for your chosen AI Employee.

  • Turn on the capabilities that are most important for your workflow and customer experience.
  • Add or adjust capabilities as your business needs evolve.
  • Use the additional instructions/goals field to give extra context or specific directions.
  • Create your own custom capabilities: You can design custom tasks for your AI Employee to perform, using the Custom Capabilities builder. This is great for advanced workflows or unique business needs. Learn more ›

Examples of capabilities you might enable:

  • Capture lead information (e.g., name, email, phone)
  • Book appointments using your connected calendar
  • Answer common customer questions (FAQs)
  • Follow your specific instructions or goals
Writing capability instructions

When adding instructions to capabilities, be specific about when and how to use them:

Good: "Only ask for contact information after the customer shows clear interest in our services. If they're just browsing, answer their questions first."

Too vague: "Get contact info when appropriate."

You can experiment with different capabilities to see what works best, and update them any time as your business grows.

Tools​

The Tools section enables your AI Employee to interact with external systems and retrieve real-time information during conversations.

What this is:
Tools are connections to APIs (external software systems) that your AI Employee can use to perform actions or retrieve information that goes beyond their built-in knowledge. For example, tools can check weather forecasts, look up inventory, create records in your CRM, or verify appointment availability in real-time.

What to do:
Tools are typically configured as part of custom capabilities. When you need your AI Employee to:

  • Access real-time information (weather, inventory, account details)
  • Create or update records in external systems
  • Automate workflows across multiple platforms
  • Make decisions based on live data

You'll create a custom capability with an integrated tool. The tool acts as a bridge, letting your AI fetch or send information to other software.

Examples of what tools can do:

  • Check weather conditions before scheduling outdoor work
  • Look up customer order history from your e-commerce platform
  • Verify product availability in your inventory system
  • Create new leads in your CRM after capturing contact information
  • Check calendar availability before booking appointments
Learning more about tools

Tools require some understanding of APIs (how software systems communicate). If you're new to this concept, start with the Tools & Integrations Overview to learn the fundamentals before building your first tool.

Most standard business needs (like appointment booking or lead capture) are handled by built-in capabilities that already have tools configured. Custom tools are for specialized integrations specific to your business.

Knowledge sources​

The Knowledge Sources section is where you teach your AI Employee about your business, so it can give accurate, up-to-date answers to customers.

What this is:
Knowledge sources are collections of business information your AI can refer to during conversations—such as website pages, FAQs, or uploaded documents. The AI uses this info only when it needs to answer a relevant question.

What to do:
Add any information that would help your AI answer questions or provide better service:

  • Add your website: Let the AI "read" your site to answer questions about your offerings, prices, or policies.
  • Upload documents or custom text: Share menus, price lists, terms, or answers to frequently asked questions.
  • Provide basic business information: Include essentials like your address, hours, services, and contact info.
  • Review and update regularly: Make sure your knowledge sources are current, so your AI never gives outdated info.
tip

The more up-to-date and specific your knowledge sources, the more helpful and accurate your AI Employee will be.

Save & test​

  • Save your configuration:
    Don't forget to save changes before leaving the page!
  • Test your AI:
    Most AI Employees let you simulate conversations or actions, so you can preview how they'll interact with real customers.

Testing and optimizing your AI Employees​

After initial setup, regularly test and optimize your AI Employees to ensure they're performing effectively.

Testing strategies​

Test Systematically

  • Ask the same question in different ways to verify consistency
  • Test what happens when customers provide incomplete information
  • Try scenarios that should trigger specific capabilities
  • Verify the AI's tone and behavior match your brand

Clear Cookies for Fresh Tests

  • For advanced testing, clear cookies between test conversations
  • This ensures each test starts without previous context
  • Helps verify the AI works correctly for new visitors

Monitoring performance​

Review Conversation Logs

  • Check both identified conversations and anonymous visitors
  • Look for patterns in successful vs. problematic interactions
  • Identify common customer questions that need better responses

Use AI Explanations

  • Click "Explanation" on any AI response to see:
    • The reasoning behind the AI's decisions
    • Which knowledge sources were used
    • What tools were called and their API responses
  • Use this information to identify areas for improvement

Iterative improvement process​

1. Identify Issues

  • Specific responses that miss the mark
  • Capabilities that trigger incorrectly or not at all
  • Knowledge gaps where the AI can't answer questions
  • Tone or behavior that doesn't match expectations

2. Determine the Root Cause

  • Purpose issue: Adjust if it affects overall behavior across all conversations
  • Capability issue: Refine when specific actions aren't working correctly
  • Knowledge issue: Update when facts are missing or incorrect
  • Tool issue: Fix when API calls fail or return wrong data

3. Make Focused Changes

  • Change only one thing at a time
  • Document what you changed and why
  • Test immediately to verify the change worked

4. Monitor Results

  • Watch real conversations for the impact of your changes
  • Verify the issue is resolved without creating new problems
  • Continue iterating based on ongoing performance
Testing best practices
  • Test after every configuration change
  • Monitor live conversations regularly (weekly is a good starting point)
  • Keep notes on what works and what doesn't
  • Be patient—optimizing AI Employees is an iterative process that improves over time