Answer
Your chatbot and voice assistant have built-in safety measures that are automatically active. These protect both your visitors and your brand. You don't need to configure anything.
What is handled automatically?
Crisis situations
When a visitor shows signs of a crisis situation (such as suicidal thoughts or self-harm), the assistant stops the regular conversation and immediately refers them to professional help services:
- Netherlands: 113 Suicide Prevention (call 113 or free 0800-0113, chat at 113.nl)
- Germany: Telefonseelsorge (0800 111 0 111 / 0800 111 0 222 / 116 123)
- International: 112 or findahelpline.com
The assistant does not give any other advice and does not ask further questions in these situations.
Illegal activities
Requests that (directly or indirectly) assist with illegal activities are always refused. This also applies to questions framed as hypothetical or educational.
Knowledge base protection
Your knowledge base contains your expertise (marketing, coaching, consulting, etc.). The assistant uses this knowledge to help visitors, but refuses requests to use that knowledge in ways that could harm or manipulate others.
How does this work technically?
The safety rules are automatically added to every conversation, both in text chat and in the voice assistant. They have the highest priority and cannot be overridden by other instructions.
This works together with the built-in safety features of the language models powering your chatbot. Together, they ensure a reliable and safe experience for your visitors.
Do I need to configure anything?
No. The safety measures are active by default for all chatbots and voice assistants on the platform. You cannot disable them (and that is by design).
Tips
- Test your chatbot regularly to see how it responds to different types of questions
- The safety measures work in Dutch, English, and German
- If you have concerns about the safety of a specific situation, you can always contact us