AI Sentiment

Use AI Sentiment when you need to understand how AI answers describe a brand, beyond whether it is mentioned.

What to provide

Provide a question and one brand. Optionally choose the AI surfaces. Keep the buyer context specific so that the tone is meaningful.

You can also specify a country for the AI measurement; Gemini has no country setting. See models and markets for defaults, the optional ChatGPT surface and measurement differences.

What you get back

  • Positive, neutral and negative tone weights per model.
  • Phrases supporting the classifications.
  • An aggregate across usable model classifications.

Example

ChatGPT
Cite42 connected
YouHow do AI models describe Acme CRM when answering “best startup CRM”? Show the positive and negative phrases behind the sentiment scores.
MCPAI Sentiment
ChatGPTThe sample OpenAI classification is 55% positive, 35% neutral and 10% negative, supported by phrases such as “modern, AI-native CRM” and “polished UI”. These are tone weights for one answer, not a survey of customers.
Illustrative conversation. Results vary by request.

Reading the results

Sentiment percentages are AI-estimated tone weights, not the percentage of answers that are positive. Failed classifications are excluded. No usable answers means unavailable, not neutral or 0%.

Billing

$0.03 to $0.08 / model. The selected models determine the total; fewer models cost less. Failed calls are not charged. See Balance & billing.

Next steps

AI Visibility: Measure how often the brand appears.

Building a direct integration? The REST reference contains exact parameters and request and response examples.

About AI Sentiment →