Goodbyes

The “Goodbyes” tracker detects phrases that signal the end of a conversation, such as polite farewells, acknowledgments, or closing statements. By capturing these mentions, teams can assess whether calls conclude satisfactorily, whether proper courtesy is maintained, and if any follow-up actions are clearly stated before parting.

Who Benefits:

• Project Managers & Team Leads: Understand if conversations end appropriately, ensuring representatives follow closing protocols and customers leave on a positive note.
• Sales & CX Teams: Confirm that calls end smoothly, with gratitude and well-wishes, improving overall satisfaction and leaving a good final impression.
• Business Managers & Strategists: Spot patterns in how calls conclude, informing training on closing techniques or identifying when more extensive follow-up instructions might be needed.
• Compliance Officers: Ensure that goodbye sequences adhere to any required disclaimers, acknowledgments, or compliance-related closing statements.

Value Proposition:

• Closure & Clarity: Highlighting farewell exchanges ensures calls end politely and customers understand any next steps.
• Efficiency: Structured JSON output feeds into dashboards, supporting analysis on conversation length, completion rates, and adherence to protocols.
• Continuous Improvement: Insights into how goodbyes are handled inform refinements to closing scripts, training materials, and final acknowledgments.

Data Dictionary

Objective:

Detect phrases in the conversation that indicate a farewell, closing remark, or polite goodbye. The tracker returns occurrences of matched phrases, along with timestamps and confidence. This data helps teams confirm that calls conclude professionally and courteously.

Proposed JSON Schema:

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

Data Dictionary:

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

{
  "trackers": []
}

Determining Confidence & Thresholds

How the Goodbyes Tracker Works:

• Seed the tracker with phrases related to farewells (e.g., "goodbye," "take care," "have a nice day," "thank you for calling," "talk to you later").
• The system interprets these phrases to find semantically similar statements in the transcript.
• Confidence indicates how closely the detected phrase relates to a conversation-ending sentiment.

Confidence & Calibration:

• Initial Testing: Validate on known calls to ensure detected phrases accurately represent a genuine farewell moment.
• Refine Phrases: Add or remove seed phrases to capture evolving closing styles or cultural nuances in farewells.
• Thresholding: Downstream consumers can filter low-confidence matches if necessary.
• Stakeholder Feedback: Involve CX leads or compliance officers to refine what constitutes a proper call closure according to brand standards.

Prompt Construction & Instructions

Role Specification & Reiteration:

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

No Hallucination:

• Only detect goodbyes 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 'Goodbyes' mentions in a conversation. You have:

• A transcript
• A 'Goodbyes' tracker defined by seed phrases indicating farewell or closing remarks.

Your task:

1. Detect phrases suggesting the call is ending politely and clearly.
2. Return text, timestamp, and confidence.
3. Return only JSON as per the schema, no extra commentary.

Instructions & Variables (example):

{
  "tracker": {
    "name": "Goodbyes",
    "phrases": [
      "thank you, have a great day",
      "goodbye",
      "take care",
      "talk to you later",
      "bye now"
    ]
  }
}

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

[TRANSCRIPT_JSON][TRANSCRIPT_JSON]"

Expected Output Format:

{
  "trackers": [
    {
      "name": "Goodbyes",
      "matches": [
        {
          "text": "Thank you for your time today, have a great day!",
          "timestamp": "2024-07-10T11:45:00Z",
          "confidence": 0.9
        }
      ]
    }
  ]
}