• Be specific about the desired outcome (e.g., “Identify quantifiable pain points…” not just “Find pain points”).
• Mention the intended audience or use case (e.g., “…to inform coaching sessions”, “…for tailoring follow-up emails”).
• Keep it concise but comprehensive (1-3 paragraphs is often sufficient).

“This field aims to extract the concrete, actionable next steps agreed upon during a sales call. It should capture who is responsible for each action item and any specific deadlines mentioned. The output will be used by reps to ensure follow-through and by managers to track deal progression.”

• It must define clear criteria for each score level (typically 0-10 or 1-10).
• Describe observable behaviors or specific content that corresponds to each level.
• Clearly differentiate between score levels. What makes an 8 different from a 6?
• Use Markdown for structure (headings for score levels, bullets for criteria).
• Example Snippet (Discovery Score Rubric):
  Copy

  ### 8/10: Strong Discovery
  
  - Rep successfully uncovered and quantified at least 2 major pain points.
  - Rep identified the primary decision-maker and their key criteria.
  - Rep established a clear understanding of the prospect's timeline and budget range.
  - Minor gaps in understanding the full technical landscape may exist.
  
  ### 6/10: Competent Discovery
  
  - Rep identified general pain points but struggled to quantify impact.
  - Rep identified some stakeholders but not clearly the primary decision-maker.
  - Basic understanding of timeline or budget, but specifics are vague.
  

• Clearly state what information to extract or what content to generate.
• Be specific about the desired format (e.g., “Output as a Markdown bulleted list”, “Generate a 3-paragraph summary”, “Extract verbatim quotes tagged with the speaker”).
• Break down complex tasks into smaller steps if necessary.
• Example (Competitor Mentions Value Field):
  Copy

  1.  Scan the transcript for mentions of competitor companies or products.
  2.  For each mention, extract:
      - The competitor's name.
      - The context of the mention (e.g., currently using, evaluating, comparing features).
      - Any specific pros or cons mentioned about the competitor.
  3.  Format the output as a Markdown list, with each competitor as a main bullet point and context/pros/cons as sub-bullets.
  

• Provide concrete examples of “good” vs. “bad”.
• Address potential ambiguities or edge cases.
• Specify desired tone or level of detail.
• Guide the AI on how to interpret specific situations mentioned in the transcript/document.

• Use clear examples directly related to the Instructions/Rubric.
• Explain why something is good or bad according to your process.
• Don’t repeat the Instructions; elaborate on them.

• Good Example (Budget): Rep asks “What budget range have you allocated for solutions like this?” or “How does your typical purchasing process work for software in this price range?”
• Poor Example (Budget): Rep avoids the budget question entirely or only asks “Do you have budget?” without probing further.
• Edge Case: If the prospect explicitly states they cannot discuss budget on this call, do not penalize the rep in the ‘Cons’ section for not getting a number, but note the lack of budget information.

• MEDDPICC: Our methodology for deal qualification (Metrics, Economic Buyer, Decision Criteria, Decision Process, Identified Pain, Champion, Competition).
• ARR: Annual Recurring Revenue.

• QBR: Quarterly Business Review.