We have currently had ~500 users since the launch last week, and the satisfaction survey shows that about 78% think the results were accurate so far (of course, the sample size is small, but I wanted to share the rough data).

Sometimes the AI still makes some silly mistakes, but I strongly believe that it will only get better.

Please let me know if you find any specific bugs or errors. I will be happy to improve it based on your valuable feedback.