Short answer
Customer satisfaction score (CSAT) is a metric that measures how satisfied customers are with a specific interaction or experience. It is usually calculated as the percentage of respondents who give a positive rating (for example a 4 or 5 on a 5-point scale) on a short survey sent right after a support conversation.CSAT is interaction-level and immediate, which makes it different from relationship metrics like NPS. It tells you whether a given ticket, chat, or call left the customer feeling helped. Because response rates are low and self-selected, it is best read as a trend over many interactions rather than a verdict on any single one.
CSAT vs NPS vs CES at a glance
| Dimension | CSAT | NPS | CES |
|---|---|---|---|
| Asks about | one interaction | the whole relationship | effort to get help |
| Scoring | percent rating 4 or 5 | promoters minus detractors | mean ease rating |
| Best read as | trend over many interactions | slow loyalty signal | friction detector |
How to calculate a CSAT score
CSAT score = positive responses ÷ total responses × 100. If 200 customers answer the survey and 168 choose 4 or 5 stars, the score is 84 percent. Count only people who answered, and report how many did, because a score built on 12 responses moves a lot from week to week. The free CSAT calculator does the sum from your counts.
CSAT when an AI agent answers
CSAT is an industry-standard metric, and it gets harder to read once an AI agent answers part of the queue. A blended score can hide a weak AI behind a strong team, or the other way round. Aide, the AI platform for customer service, sends a 1 to 5 star survey in Aide Chat and reports CSAT for the Aide AI Agent and for the team separately, with the rating spread, trend and latest comments. Read it alongside resolution rate rather than chasing it on its own: satisfaction follows from fixing the problem.