Dissatisfaction

The “Dissatisfaction” tracker detects phrases that signal negative sentiment—such as frustration, disappointment, or unhappiness—from customers or participants. By capturing these mentions, teams gain insights into what goes wrong, where improvements are needed, and how to respond proactively to enhance satisfaction and retention.

Who Benefits:

• Project Managers & Team Leads: Identify common triggers of dissatisfaction, guiding process improvements and tailored training sessions.
• Sales & CX Teams: Recognize dissatisfaction cues early, enabling timely interventions and empathetic responses to address complaints effectively.
• Business Managers & Strategists: Spot trends in negative feedback, informing product refinements, policy adjustments, or shifts in customer engagement strategies.
• Compliance Officers: Ensure dissatisfaction is acknowledged and handled according to guidelines, preventing escalation and maintaining trust.

Value Proposition:

• Actionable Insights: Highlighting dissatisfaction moments enables teams to take swift corrective measures, improving customer experiences.
• Efficiency & Scalability: Structured JSON output seamlessly integrates with dashboards for trend analysis and strategic decision-making.
• Continuous Improvement: Understanding recurring dissatisfaction themes guides iterative enhancements to products, services, and processes.

Data Dictionary

Objective:

Detect phrases in the conversation that express dissatisfaction, disappointment, or frustration. Rather than a numeric score, the tracker returns occurrences of matched phrases, along with timestamps and confidence. This data helps teams pinpoint and address negative sentiment effectively.

Proposed JSON Schema:

{
  "trackers": [
    {
      "name": "Dissatisfaction",
      "matches": [
        {
          "text": "<string>",
          "timestamp": "<ISO 8601 or relative time>",
          "confidence": <number> // 0.0 to 1.0
        }
      ]
    }
  ]
}

Data Dictionary:

If No Matches Found:
If no dissatisfaction references are detected, return an empty matches array for "Dissatisfaction" or omit the tracker. If no trackers match at all, return:

{
  "trackers": []
}

Determining Confidence & Thresholds

How the Dissatisfaction Tracker Works:

• Seed the tracker with phrases representing negative sentiment (e.g., "unhappy," "frustrated," "not satisfied," "this is terrible").
• The system interprets these phrases and identifies semantically similar statements in the transcript.
• Confidence indicates how strongly the detected phrase conveys dissatisfaction.

Confidence & Calibration:

• Initial Testing: Validate on known calls to ensure detected phrases accurately reflect negative sentiment.
• Refine Phrases: Add or remove seed phrases to capture evolving language or subtle variations of dissatisfaction.
• Thresholding: Downstream consumers can filter out low-confidence matches if needed.
• Stakeholder Feedback: Involve CX leads or compliance teams to refine definitions and ensure relevant dissatisfaction cues are captured.

Prompt Construction & Instructions

Role Specification & Reiteration:

• Present the system as a “highly experienced assistant” identifying dissatisfaction mentions.
• Reiterate instructions: return only JSON, no extra commentary, follow schema.

No Hallucination:

• Only detect dissatisfaction references if the transcript supports it.
• If uncertain, assign lower confidence or omit matches.

Strict Formatting:

• Return only JSON with trackers.
• If none found, return empty trackers.

Example Prompt for Implementation

System Message (Role: System):
"You are a highly experienced assistant that identifies 'Dissatisfaction' mentions in a conversation. You have:

• A transcript
• A 'Dissatisfaction' tracker defined by seed phrases indicating negative sentiment.

Your task:

1. Detect phrases showing customer disappointment or frustration.
2. Return text, timestamp, and confidence.
3. Return only JSON as per the schema, no extra commentary.

Instructions & Variables (example):

{
  "tracker": {
    "name": "Dissatisfaction",
    "phrases": [
      "not happy",
      "unhappy",
      "frustrated",
      "this is terrible",
      "disappointed"
    ]
  }
}

User Message (Role: User):
"Analyze the following transcript and detect 'Dissatisfaction' mentions:

[TRANSCRIPT_JSON][TRANSCRIPT_JSON]"

Expected Output Format:

{
  "trackers": [
    {
      "name": "Dissatisfaction",
      "matches": [
        {
          "text": "I'm very disappointed with how this turned out.",
          "timestamp": "2024-07-10T09:40:00Z",
          "confidence": 0.9
        }
      ]
    }
  ]
}