Satisfaction Signals

What are Satisfaction Signals?

Satisfaction signals are indicators of how users feel about their AI interactions. Thunder automatically detects two types:

CSAT (Satisfaction) - Signals that the user is happy, their question was answered, or the AI was helpful
DSAT (Dissatisfaction) - Signals that the user is frustrated, confused, or the AI failed to help

How Signals are Detected

Thunder analyzes conversation content to identify explicit and implicit signals: CSAT Examples:

“Thanks, that’s exactly what I needed!”
“Perfect, you’re a lifesaver”
“This worked great”

DSAT Examples:

“That’s not what I asked”
“You’re not understanding me”
“This is frustrating”
“Let me talk to a human”

Detection uses AI models trained on conversational patterns - no explicit feedback buttons required.

Signal Metrics

Query signals to understand their frequency:

Metric	Description
`sessions`	Conversations where this signal was detected
`users`	Unique users who expressed this signal
`messages`	Number of times this signal appeared

Querying Signals

Get all satisfaction signals with their frequency:

curl "https://api.usethunder.com/v1/query?\
groupBy=CSAT_DSAT&\
instanceId=your-instance-id&\
dateRange.start=2024-01-01T00:00:00Z&\
dateRange.end=2024-01-31T23:59:59Z&\
timeGranularity=ALL&\
fields=sessions,users,messages&\
sort.field=sessions&\
sort.order=DESC" \
  -H "x-api-key: YOUR_API_KEY"

Response:

{
  "elements": [
    {
      "dimensions": {
        "csatDsatSignal": {
          "id": "signal-uuid",
          "name": "Expressed gratitude",
          "signalType": "SAT",
          "description": "User thanked the assistant"
        }
      },
      "fields": {
        "sessions": 523,
        "users": 412,
        "messages": 687
      }
    },
    {
      "dimensions": {
        "csatDsatSignal": {
          "id": "signal-uuid-2",
          "name": "Expressed confusion",
          "signalType": "DSAT",
          "description": "User indicated they didn't understand"
        }
      },
      "fields": {
        "sessions": 89,
        "users": 76,
        "messages": 124
      }
    }
  ]
}

Session-Level Metrics

When querying sessions, you get aggregate signal counts:

curl "https://api.usethunder.com/v1/query?\
groupBy=SESSION&\
instanceId=your-instance-id&\
dateRange.start=2024-01-01T00:00:00Z&\
dateRange.end=2024-01-31T23:59:59Z&\
timeGranularity=ALL&\
fields=messages,sat,dsat,netSat" \
  -H "x-api-key: YOUR_API_KEY"

Field	Description
`sat`	Count of CSAT signals in the session
`dsat`	Count of DSAT signals in the session
`netSat`	Net satisfaction score: `(sat - dsat) / messages * 100`

Finding Problem Sessions

Filter for sessions with specific signals:

curl "https://api.usethunder.com/v1/query?\
groupBy=SESSION&\
instanceId=your-instance-id&\
signalIds=dsat-signal-uuid&\
dateRange.start=2024-01-01T00:00:00Z&\
dateRange.end=2024-01-31T23:59:59Z&\
timeGranularity=ALL&\
fields=messages,sat,dsat&\
dimensionFields=title,summary" \
  -H "x-api-key: YOUR_API_KEY"

This returns only sessions where that specific DSAT signal was detected.

Use Cases

Track Overall Satisfaction

Monitor the ratio of CSAT to DSAT signals over time to gauge AI quality.

Identify Failure Patterns

Examine sessions with high DSAT to understand common failure modes.

Measure Improvements

Track signal trends before and after AI updates to measure impact.

Prioritize Fixes

Focus on the most common DSAT signals first for maximum impact.

See Querying Metrics for more query patterns and the full field reference.

Getting Started

Fundamentals

Features

Guides

Security

Satisfaction Signals

What are Satisfaction Signals?

How Signals are Detected

Signal Metrics

Querying Signals

Session-Level Metrics

Finding Problem Sessions

Use Cases

Track Overall Satisfaction

Identify Failure Patterns

Measure Improvements

Prioritize Fixes

​What are Satisfaction Signals?

​How Signals are Detected

​Signal Metrics

​Querying Signals

​Session-Level Metrics

​Finding Problem Sessions

​Use Cases

Track Overall Satisfaction

Identify Failure Patterns

Measure Improvements

Prioritize Fixes

What are Satisfaction Signals?

How Signals are Detected

Signal Metrics

Querying Signals

Session-Level Metrics

Finding Problem Sessions

Use Cases