I think the visual search would work better if it takes a picture when you press the button and then you can take your time speaking your prompt.
2 Likes
Hi @Krayzielilsmoki ,
I agree with your idea.
In my opinion an other sequence should be choosen, too.
For example: First a picture is taken from a document. Then you can see if the quality or focus of the picture will fit to your next prompt.
Second you can easily concentrate on your own prompt.
At the moment I can speak my prompt with the chance in a negative way, that my picture can not focus all relevant details of an object/document etc.
1 Like
Really interesting point. 
I don’t think there’s much that could go in the wrong direction. But would it really reduce the risk of taking bad snapshots?