BrandMentions AI works best when humans stay involved. Your feedback helps the AI become more accurate, less biased, and more useful for you.
Whenever you see an AI insight that is especially helpful or clearly off, you can let the system know directly in the platform. This creates a human in the loop feedback cycle that improves results over time.
1. Correcting AI insights inside your project
The fastest way to “teach” the system and fix your own data is to correct what you see directly in the interface.
Correcting wrong sentiment on mentions
If a mention is clearly labeled with the wrong sentiment:
Open the Mentions feed.
Find the mention with the incorrect sentiment label.
Change the sentiment manually to the correct value, for example from positive to negative.
Validating AI Analysis and AI Assistant results
When an AI Analysis summary or an AI Assistant answer looks suspicious:
Click through to the underlying mentions or charts.
Confirm what is actually happening in the raw data.
Note what feels wrong, for example:
“The summary says sentiment dropped, but the chart is stable.”
“This answer treats sarcasm as positive feedback.”
You will use these notes when you contact the BrandMentions team.
2. How to report inaccurate AI insights to BrandMentions
If an AI insight is clearly wrong, biased, or repeatedly unhelpful, you can report it so the team can investigate and improve the models.
What you can send to the BrandMentions team
When you contact support (for example via in app chat or email), include:
A short description of the issue
“AI Analysis says there was a spike in negative sentiment, but mentions are actually positive.”
Where it happened
Project name
Section (Mentions feed, specific chart, AI Assistant, etc.)
Screenshots, if possible
The AI output that seems wrong
The mentions or chart that prove it is incorrect
This kind of detailed example is extremely helpful for debugging and model improvement.
Why your feedback is important
Your feedback:
Improves the AI models
Real examples of mistakes are some of the best training signals for future improvements.
Helps detect and reduce bias
If you see patterns that feel unfair or skewed, reporting them helps the team correct them.
Makes the product better for you
Over time, the AI becomes more aligned with your brand language, your audience, and your use cases.
By taking a moment to correct data and share clear examples with the BrandMentions team, you actively help make AI insights smarter, more accurate, and more trustworthy for your own work and for every other user.