Promotion Mention

The “Promotion Mention” tracker detects instances where a representative references promotional offers—such as discounts, coupon codes, or special pricing. Insights from these mentions inform sales effectiveness, marketing success, and help ensure representatives follow approved promotional strategies.

Who Benefits:

• Project Managers & Team Leads: Understand how often representatives mention promotions, guiding training and identifying opportunities for improvement.
• Sales & CX Teams: Assess if agents align with current promotional campaigns, ensuring customers are aware of available deals.
• Business Managers & Strategists: Track the prevalence and impact of promotions, feeding data into marketing and product optimization decisions.
• Compliance Officers: Confirm that only approved promotions are mentioned and that agents follow policy in communicating special offers.

Value Proposition:

• Sales Optimization: Identifying promotion mentions helps measure campaign reach and agent compliance with promotional guidelines.
• Customer Engagement: Ensure customers are informed about relevant deals, enhancing satisfaction and conversion rates.
• Operational Efficiency: Structured JSON output integrates with dashboards, supporting analytics and strategic decision-making.

Data Dictionary

Objective:

Detect phrases related to promotions, coupons, discounts, or special offers. Rather than a numeric score, the tracker returns occurrences of matched phrases along with timestamps and confidence. This data helps you understand when and how often promotions are introduced in calls.

Proposed JSON Schema:

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

Data Dictionary:

If No Matches Found:
If no promotional mentions are detected, return an empty matches array for the "Promotion_Mention" tracker or omit the tracker. If no trackers match, return:

{
  "trackers": []
}

Determining Confidence & Thresholds

How the Promotion_Mention Tracker Works:

• Define seed phrases related to promotions (e.g., "special offer," "discount code," "coupon," "limited-time deal").
• The system interprets these phrases and finds semantically similar mentions in the transcript.
• Confidence indicates how clearly the phrase suggests a promotion.

Confidence & Calibration:

• Initial Testing: Validate on known calls to ensure detected phrases accurately represent promotions.
• Refine Phrases: Add or remove seed phrases to capture evolving promotional language.
• Thresholding: Downstream consumers can ignore low-confidence matches.
• Stakeholder Feedback: Integrate insights from marketing or compliance teams to refine what qualifies as a promotion mention.

Prompt Construction & Instructions

Role Specification & Reiteration:

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

No Hallucination:

• Only detect promotional mentions if evidenced by the transcript.
• If uncertain, assign lower confidence or omit the match.

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 'Promotion_Mention' in a conversation. You have:

• A transcript
• A 'Promotion_Mention' tracker defined by seed phrases indicating promotions, discounts, or special offers.

Your task:

1. Detect phrases suggesting a promotion is offered.
2. Return text, timestamp, and confidence.
3. Return only JSON as defined. No extra commentary.

Instructions & Variables (example):

{
  "tracker": {
    "name": "Promotion_Mention",
    "phrases": [
      "special discount",
      "coupon code",
      "promotional offer",
      "20% off",
      "limited-time deal"
    ]
  }
}

User Message (Role: User):
"Analyze the following transcript and detect 'Promotion_Mention' occurrences:

[TRANSCRIPT_JSON][TRANSCRIPT_JSON]"

Expected Output Format:

{
  "trackers": [
    {
      "name": "Promotion_Mention",
      "matches": [
        {
          "text": "We’re currently offering a 20% discount if you sign up today.",
          "timestamp": "2024-07-10T10:45:00Z",
          "confidence": 0.9
        }
      ]
    }
  ]
}