Skip to main content

Feedback

Feedback, it's important. Really.

The Carrot and the StickYour feedback when using the AI Assistant is very valuable! Not only does it let us know what you think, but it also helps by improving the quality of our personas and models for the future. So by giving feedback, you will get better responses in the future.

Chat buttonsThere are two ways of giving feedback, either by using the thumbs-up button or by using the thumbs-down button. You can find both at the top right of each answer.

Giving positive feedback

If you get a really good or helpful answer, let us know! It's not only important to know when an answer was not satisfactory, but also when it was good. This reinforces our AI models when it has performed well, so it will tend to give more similar responses in the future. It's like giving your favorite Noodle a pat on the back!

Giving negative feedback

Sometimes an answer might not live up to your expectations. When this is the case, you can let us know by clicking the thumbs-down button. It is very important that you let our Noodles know what went wrong. When you click the button you can select why the answer was not as expected. The categories are:

  • Inaccurate: the answer was not (fully) correct.
  • Out of date: the answer did not contain up-to-date information.
  • Harmful or offensive: the answer was in some way harmful to you or others, inappropriate or offensive. Minorities may have been insulted.
  • Refused when it shouldn’t have: the AI refused to answer your question. Certain persona’s will not react when the question is off-topic or unrelated to it’s objective. If you are in this situation, but you think it should have answered, select this option.
  • Didn’t fully follow instructions: the AI did not (completely) followed the instructions in your question. Like when you have children, they do that too.
  • Other: well, none of the above I guess.

In addition to this please explain in a few words what exactly was wrong about the answer, and what you think it should have been, or how it could have been better.

Negative feedback dialog

What do we do with your feedback?

First and foremost, your feedback is used to improve our personas and models so they can better serve you in the future. To do this, we store the conversation, along with your name and email address. The conversation will be used to re-train our models and/or adapt our personas. Your name and email will be used if we need to contact you for further information or details. Please do not submit your feedback if you feel that your chat contains personal data or sensitive information. The feedback will be processed both automatically, as well as read by our Noodle-handlers (which are human, by the way) for evaluation.

When the feedback has been processed, we will not keep your chat and feedback any longer as a separate dataset, but please note that it can be used to re-train the model.

If you want to know more about what we do with your data, you can find it here: Yourprivacy Data, Our Care: How We Safeguard Your Informationstatement.