“...support team resolved it in under an hour, genuinely impressed...”
Sentiment Analysis
Know how people feel, and about what.
Every mention is scored positive, neutral or negative the moment it arrives, then grouped into topics like pricing, support and product quality. An average tells you something changed. Topics tell you why.
How scoring works
Read in context, not by keyword.
The model reads the whole mention rather than counting positive or negative words. That is how a sarcastic 'great, another outage' ends up negative instead of positive.
MentionSentimentTopicWhy
“...support team resolved it in under an hour, genuinely impressed...”
PositiveSupportResolution + praise“...does what it says. Nothing more, nothing less...”
NeutralProduct qualityFactual, no strong tone“...renewal price jumped without any heads up...”
NegativePricingComplaint about a change“...great, another outage right before a deadline...”
NegativeReliabilitySarcasm read in context
Topic clustering
The average hides the story. Topics reveal it.
Overall sentiment can look stable while one topic quietly turns negative. In the sample below, pricing carries almost half of all negative mentions even though the brand is 71% positive overall.
Topics are created from recurring themes in your own mentions
Each topic has its own sentiment split and trend line
Rename, merge or pin topics that matter to your team
Topic breakdown · last 30 days
Sample 71% positive 17% neutral 12% negative
- Support412 mentions
- Pricing286 mentions
- Product quality251 mentions
- Delivery174 mentions
- Onboarding96 mentions
Details
Scoring you can trust and inspect.
Three clear states
Positive, neutral and negative. Simple enough to report on, precise enough to act on.
Automatic topics
Mentions are grouped into themes without you building keyword rules for each one.
Trend over time
Follow sentiment for the brand and each topic by day, week or month.
See the evidence
Click any score or topic to read the exact mentions behind it.
Correct and retag
Change a mention's sentiment or topic when your team knows better. Reports update to match.
Source-aware reading
A one-line social post and a long review are read differently, because people write differently in each.
Questions
Sentiment FAQ
How accurate is automated sentiment scoring?
Automated scoring is strong on clear opinions and weaker on heavy sarcasm, mixed opinions and very short posts. That is why every score links to the original mention and can be corrected by your team.
Can one mention have more than one topic?
A mention is assigned to its main topic for clean reporting. Mentions that clearly cover more than one theme can be retagged by your team.
Do corrections change historical reports?
Yes. When a mention's sentiment or topic is corrected, dashboards and future digests reflect the change.
Is sentiment the same as star rating?
No. A three-star review can be warm or frustrated. Kenaura reads the text itself, and shows the star rating alongside when a review site provides one.
Keep exploring
The rest of the monitoring loop
Listening now
Your next reputation shift is already being written.
Point Kenaura at your brand and the first mentions start flowing into your feed, scored and grouped, the same day.