What is CSAT (customer satisfaction score)?
CSAT measures how happy customers were with a specific interaction — a support chat, a delivery, an onboarding call. You ask one question (“How satisfied were you with…?”), collect ratings on a scale (usually 1–5), then turn the top responses into a percentage. That percentage is your customer satisfaction score — simple to explain, easy to track week over week.
How to read every box in the CSAT calculator
Meet Alex, who runs support for a growing ecommerce brand. After every resolved ticket, customers get a 1–5 survey. Last month 200 people responded. Alex plugs the counts into this CSAT calculator before the team standup.
Alex's post-support survey
Scores 1–2: 35 unsatisfied · Score 3: 40 neutral · Scores 4–5: 125 satisfied. CSAT = 125 ÷ 200 × 100 = 62.5%. Average rating = 3.63 out of 5 — useful context, but not the same number as CSAT.
Inputs — what you type in
- Rating scale (1–5, 1–7, or 1–10)
- Most CSAT surveys use 1–5. Larger scales change which scores count as “satisfied” — the calculator adjusts automatically.
- Top 2 box vs top box only
- Standard CSAT uses top 2 box: ratings 4 and 5 on a 5-point scale (SurveyMonkey). Top box only (5s) is stricter — good for high-stakes interactions.
- Satisfied / neutral / unsatisfied counts
- Group mode for quick entry. Alex has 125 / 40 / 35. Or enter each score individually on the “By score” tab.
- What-if: win back unsatisfied (%)
- Models support improvements — if 20% of the 35 unsatisfied customers (7 people) move to satisfied, CSAT rises from 62.5% → 66%.
Outputs — what you get back
- CSAT score (%)
- The headline: share of satisfied responses. Alex's 62.5% means roughly three in five customers left happy — room to improve.
- Average rating
- Mean score out of 5 (or your scale). Helpful for trends; do not confuse it with CSAT percentage.
- Projected CSAT & charts
- Score distribution, satisfied/neutral/unsatisfied mix, and how CSAT climbs as you recover unhappy respondents.
The problem: one bad support chat shows up here
NPS asks “would you recommend us?” CSAT asks “was that interaction good?” That makes CSAT the right metric after support tickets, live chats, and deliveries — exactly where slow or wrong answers hurt. A single 1-star survey is a data point; a cluster of them is a product or staffing signal.
The most common mistake is counting only 5s as satisfied when your benchmark database uses top 2 box, or excluding neutral 3s from the denominator — both inflate the score (Kicue). This calculator keeps the math honest: all respondents in the denominator, satisfied threshold defined up front.
The CSAT formula
On a standard 5-point scale with top 2 box:
CSAT (%) = (Responses rated 4 or 5 ÷ Total responses) × 100
Example from Help Scout's worked math: 90 satisfied responses out of 120 total → 75% CSAT (Help Scout). Zendesk notes that 88% of high-performing companies use CSAT as a primary satisfaction measure — because it ties directly to specific fixes (industry statistics roundup).
“CSAT is calculated as: (Number of satisfied responses / Total responses) × 100. On a standard 5-point scale, satisfied responses are ratings of 4 and 5.”
How AI customer service moves CSAT — with numbers
CSAT after support is mostly about speed, accuracy, and whether the customer felt heard. That is where an AI agent trained on your policies changes the distribution of 4s and 5s.
- Freshworks 2024 benchmark: businesses using AI support tools saw up to 6.2% CSAT improvement, 37.6% faster first response, and 38.7% faster resolution (Freshworks Customer Service Benchmark Report).
- Glassix ecommerce study: brands using AI chatbots on messaging channels reported a 12% CSAT jump year over year (Glassix).
- Vodafone: GenAI on virtual agents drove a 20% NPS lift and up to 50% better first-time resolution on billing — the same tickets that produce 1s and 2s on CSAT surveys when they go wrong (CX Today).
Map it to Alex: 62.5% CSAT with 35 unsatisfied. Recover 20% (7 people) → 66% CSAT. A 6-point Freshworks-scale lift on 62.5% would land near 66–69% — use the what-if slider to set your own target, then measure post-launch on the same survey.
How ChatInCart helps you raise CSAT
ChatInCart is an AI customer service agent you train on your website and docs, then embed wherever customers ask for help. It answers in seconds — order status, returns, setup — instead of leaving people in a queue that turns 4s into 2s.
That shows up in CSAT drivers:
- First response time — the metric Freshworks ties directly to CSAT gains; instant chat beats “we'll email you in 24 hours.”
- Consistent answers — same policy on your site, in chat, and in email; fewer “that agent told me something else” unsatisfied scores.
- Human handoff when needed — complex tickets escalate with context so agents fix issues on the first try (first-contact resolution lifts CSAT).
Pair CSAT with our NPS calculator (loyalty) and LTV calculator (retention dollars). Start on the free plan, survey the same cohort before and after launch, and watch the satisfied column grow.
Common questions
On a 5-point top-2-box scale, many industries cluster around 75–85%. Above 90% is excellent; below 70% signals a support or product problem. Always report your scale and compare your own trend over time.
No. CSAT is the percentage of satisfied responses (top 2 box on a 5-point scale). A 4.2 average can hide a long tail of 1s and 2s — CSAT exposes that.
Right after the interaction you want to measure — ticket closed, order delivered, onboarding step completed. Transactional CSAT tells you if that specific touchpoint worked.
Yes — Freshworks benchmark data shows up to 6.2% CSAT improvement with AI-assisted support, and faster first response (up to 37%) directly affects satisfaction scores after support chats.