Skip to content
All updates

A smarter assistant

Improvement

The assistant got a series of improvements that make it better at answering complex questions about your monitors, even with large volumes of data or monitors with no tags set up.

What changed

Broad questions answered accurately. Before, questions like "what are the main topics?" on monitors with thousands of comments could lead to incomplete answers. Now the assistant processes the data directly on the server and delivers a synthesized analysis, without getting stuck halfway.

Topics described, not listed. Instead of returning raw word counts, the assistant now groups related terms and writes about each topic in natural language, including spellings with and without accents, like "vídeo" and "video", which used to be counted separately.

Post rankings. You can ask, for example, which posts got the most negative comments, the most engagement or the most likes, and the assistant can rank them and answer directly.

TikTok and YouTube in the analyses. The assistant now takes posts from these platforms into account when answering questions about posts and engagement.

Partial answer notice. When the assistant can't fully complete an analysis, it now shows a visible notice instead of ending the answer without explanation.