How sentiment works

Being mentioned by AI isn't automatically good. Locuvox reads every mention and grades how your brand is being talked about.

The five sentiment grades

Each mention of your brand is classified into one of five grades:

• Advocating — the answer actively recommends you.
• Positive — favourable, but not a direct recommendation.
• Neutral — listed or described without a slant.
• Cautious — hedged, with caveats or reservations.
• Negative — unfavourable.

Locuvox also records the answer's stance — whether it's authoritative, comparative, hedged, informational or cautionary — and any topics tied to your brand.

The seven sentiment dashboards

• Brand Perception — a quadrant map of every brand by visibility against sentiment.
• Consensus — per prompt, whether the models agree about you or split.
• Sentiment Drift — your positive-sentiment rate over time, per model.
• Model Comparison — each engine's stance, sentiment and topic profile.
• Topic Landscape — the subjects each model associates with you.
• Memory vs Retrieval — per engine, whether answers cite live sources or come from the model’s own memory.
• Ad Share of Voice — which brands buy sponsored placement on ChatGPT and Google AI Mode for your prompts, and your paid share against competitors.

Reading it well

Watch for a high mention count paired with weak sentiment — that's AI talking about you without recommending you. And check Consensus: if one model is divergent, your work is model-specific, not category-wide.

A drop in Advocating mentions is an earlier warning sign than a drop in visibility — sentiment usually moves first.

Related

• Sentiment Analysis feature
• Set up alerts & reports