Establishing clear conversation guidelines is the most effective way to ensure your AI Agent provides a high-quality, consistent experience. These guidelines act as a manual for the agent's brain, dictating how it represents your brand and interacts with visitors.
Navigating to Conversation Guidelines
- Click AI Agents from the left navigation menu.
- Select the AI Agent you want to provide guidelines for.
- Select the Conversation guidelines section.
Default Conversation Guidelines
To help you get started quickly, we automatically generate a baseline set of Conversation guidelines based on the Goal you provided when you first created your AI Agent. While these generated guidelines provide a strong foundation, you can customize the guidelines as desired to better align with your specific brand needs.
We do however recommend that your Conversation guidelines always include the following sections:
- Tone: The personality and voice used to represent your brand.
- Goals: The primary objectives guiding the AI Agent toward successful outcomes.
- Restrictions: Hard rules and boundaries to keep the AI Agent compliant.
1. Tone: The Voice of Your Brand
Your Conversation guidelines should start with a defined personality. This ensures the AI Agent is an extension of your existing brand identity.
- Personality Attributes: Select three to five adjectives that describe the voice of the AI Agent(e.g., Professional, Empathetic, Witty, Direct).
- Adaptability: Instruct the agent on how to adjust its tone based on the visitor's sentiment. For example: "Be cheerful during the initial engagement, but switch to a neutral, solution-oriented tone if a visitor expresses frustration."
- Style Standards: Define whether the agent should use industry-specific jargon, or concise bullet points.
2. Goals: Defining Success
Every interaction must have a clear objective. Without a defined goal in your Conversation guidelines, the AI Agent may provide information without actually driving value..
Provide specific goals the AI Agent should aim for, such as:
- Education Oriented: “Educate visitors on the platform’s capabilities and best practices.”
- Sales Oriented: “Encourage visitors to explore or trial a specific solution.”
3. Restrictions: Establishing Guardrails
Restrictions are the hard rules within your Conversation guidelines that keep the AI Agent safe, compliant, and focused.
- Topic Boundaries: List subjects the AI Agent must avoid (e.g., providing legal/medical advice, discussing competitor pricing, or sharing personal opinions).
- Operational Limits: Clearly state what the AI Agent cannot do, such as promising custom discounts to visitors.
- Company Policies: Specify if the AI Agent can disclose specific company policies or if certain internal documentation is off-limits.
How to Publish the Changes to Your AI Agent
Changes are automatically saved. To make the changes you have made public, click Publish changes in the top right.
For more information, please take a look at our How to Test and Publish the AI Agent section.
