Agent Disclaimers
Disclaimers allow you to display important information to users before they interact with your agent. Whether you need to meet legal requirements, set expectations about what the AI can and cannot do, or gather explicit consent, disclaimers ensure users understand key information before they begin.
What are Disclaimers?
Disclaimers are configurable messages shown to users before they start a conversation. Users must acknowledge the disclaimer to proceed, making them useful for terms of service acceptance, privacy notices, or important warnings about the agent's limitations.
Why Use Disclaimers?
Legal compliance often requires explicit user acknowledgment. Privacy notices, data processing consent, and terms of service can all be presented through disclaimers to create a clear record of user acceptance.
Setting expectations helps users understand what to expect from your agent. Explaining what the agent can and cannot do, its limitations, and the scope of support it provides reduces frustration and misuse.
Risk management is especially important for agents dealing with sensitive topics. Medical, legal, and financial applications often need explicit warnings about the limitations of AI-generated advice.
Configuring Disclaimers
To add a disclaimer to your agent, open the agent settings and find the Disclaimer section. Enable the disclaimer option, then write your disclaimer text—keep it clear, concise, and in plain language while covering all essential points.
You can configure the title shown at the top of the disclaimer, the main content text, the button label users click to acknowledge (like "I Understand" or "Accept"), and whether acknowledgment is required to proceed.
Writing Effective Disclaimers
A good disclaimer covers four key areas.
Start with what the agent does—its purpose and intended capabilities. Users should immediately understand what they're about to interact with.
Next, address limitations honestly. What can't the agent do? How accurate are its responses? What shouldn't users rely on it for?
Explain user responsibilities: how to use the agent appropriately and what to expect from the interaction.
Finally, include any necessary legal or compliance information like privacy statements or references to your terms of service.
Example Disclaimers
For a general AI assistant, you might write:
This AI assistant is designed to help with [purpose].
Please note that responses are AI-generated and may not always be accurate, this is not a substitute for professional advice, and you should not share sensitive personal information.
By continuing, you acknowledge these limitations.
For a business tool, something like:
Welcome to [Agent Name]
This tool helps [specific purpose]. Please be aware that information is for internal use only, you should verify important details before acting, and data is processed according to our privacy policy.
Click "I Understand" to continue.
For customer-facing agents:
Before we begin:
This AI assistant can help with [topics]. For complex issues, our human team is available. Your conversation may be reviewed to improve our service.
By proceeding, you accept our Terms of Service.
User Experience
When users start a conversation with an agent that has a disclaimer enabled, a modal appears displaying your disclaimer content. The user reads it and clicks the acknowledgment button to proceed. Only then does the conversation begin.
Keep disclaimers readable with clear formatting and appropriate length. Use simple language that everyone can understand. Be transparent about limitations and data use—no hidden surprises. Minimize friction by including only essential information and keeping it quick to read.
- Quick to read
- Easy to acknowledge
Use Cases
For healthcare applications, always clarify that the AI provides general health information only, is not a substitute for professional medical advice, and that users should consult healthcare providers for medical concerns.
For financial tools, note that information is for educational purposes only, does not constitute financial, tax, or investment advice, and that users should consult qualified professionals for financial decisions.
For legal applications, explain that the tool provides general legal information but does not constitute legal advice, and that users should consult licensed attorneys for legal matters.
For internal business tools, note that the tool is for authorized employees only, that usage may be logged and monitored, and that company policies apply.
Best Practices
Be specific to your actual use case rather than using generic boilerplate. Address the real limitations and requirements of your specific agent.
Be honest—don't hide limitations or bury important information. Users who feel misled become frustrated users.
Be concise. Respect users' time by including only what truly matters. Long disclaimers get skimmed or ignored.
For regulated industries, consult legal counsel to ensure your disclaimer meets requirements. Keep it updated as regulations change, and maintain records of disclaimer versions and changes.
Test that your disclaimer displays correctly across different devices and browsers. Get feedback from team members and iterate based on user experience.
Integration with Features
When sharing agents publicly through public tokens, disclaimers become especially important. External users need clear expectations, and well-crafted disclaimers provide legal protection when your agents reach people outside your organization.
For embedded agents in chat widgets, remember that space may be constrained. You might need a more condensed version of your disclaimer that still covers essential points.
Troubleshooting
If your disclaimer isn't appearing, verify that it's enabled in the agent settings and that your configuration saved properly. Try clearing your browser cache and testing again.
If text displays incorrectly, check for special characters that might cause formatting issues. Review the formatting syntax and test across different devices.
If users complain about the disclaimer, review its length and clarity. Consider condensing without losing essential information, and make sure it's not more intrusive than necessary.