Write useful Call Scorecard instructions
Last updated: October 4, 2026
What this does
Define what earns a score and ask for feedback that gives you something specific to practice on your next call.
Before you start
You need Owner or Admin access to edit a Call Type. Open Call Types, choose the type you want to edit, and enable its Call Scorecard. Write your rubric in Call Scorecard Instructions, or Scoring instructions if that label appears.
The Scorecard evaluates how you conducted the current call. It uses the transcript and your rubric, with context about the call, but not Client Memory, prior Scorecards, previous Call Notes, or CRM history.
Steps
1. Choose behaviors you can recognize in a call
Start with the few skills that matter for this Call Type. For a coaching check-in, these might include exploring obstacles and agreeing on a manageable next step; for an advisor call, they might include discovery and connecting an offer to stated needs.
Keep each criterion distinct. "Closing" and "Securing a next step" may grade the same behavior twice. Describe what the person should say or do, rather than asking Call-E to judge personality or motivation.
| Too vague or misleading | More useful instructions |
|---|---|
| "Grade discovery and give helpful feedback." | "Evaluate whether the rep explored the prospect's goal and main obstacle. Explain the score with a specific moment from the call and suggest a useful follow-up question." |
| "Give a high score if the customer buys." | "Evaluate whether the rep made an appropriate ask and clarified the decision or next step. Do not base the score on whether the prospect bought." |
| "Judge confidence and rapport." | "Evaluate whether the coach acknowledged the client's concern, asked a relevant follow-up question, and checked their understanding." |
2. Explain what the scores mean
For each criterion, describe weak, partial, and strong execution. Use one scale consistently across your rubric and explain the behavior behind its numbers.
For example, a 1-to-5 Commitment Ask criterion could use:
- 1: No clear next step was requested.
- 3: A next step was proposed, but agreement, ownership, or timing remained unclear.
- 5: A clear, appropriate ask led to an agreed next step with ownership and timing, or the prospect explicitly declined after the ask.
Use 2 and 4 for performance between those descriptions. These are example standards to adapt to your process; a purchase is not required for strong execution.
Choose criteria that belong on these calls. Call-E currently treats insufficient evidence as a low score, typically 1, and explains the lack of evidence. Asking for "N/A" does not reliably exclude a criterion, so be careful with criteria such as objection handling on calls where no objection arises.
3. Describe the feedback you want
"Give actionable advice" leaves too much open. Ask for a specific observation and one practical change, with suggested wording where it helps.
For example, imagine a rep offers to send case studies but never agrees on a follow-up:
- Too vague: "Decent close, but you need more urgency."
- Useful: "You offered case studies but left the next step open. Agree on a follow-up time before ending the call. Suggested wording: 'When would be a good time to reconnect after you've reviewed them?'"
Label invented example phrases as suggested wording, not quotations from the call. Ask Call-E to identify what to repeat when execution was strong, rather than inventing a weakness.
Keep explanations short enough to act on. If you request bold text, ask for emphasis on a specific action, such as "agree on a follow-up time," rather than a generic verdict such as "good effort."
4. Adapt this starter
This small rubric is for an advisor call. Add or change criteria to match the skills your team practices.
Evaluate the rep's behavior in this call using these two scored criteria, each from 1 to 5:
Discovery Depth: 1 means little meaningful exploration. 3 means the rep identified the goal and obstacle but left important details unexplored. 5 means the rep explored the goal, obstacle, and their impact, then confirmed their understanding.
Commitment Ask: 1 means no clear ask. 3 means a next step was proposed but agreement, ownership, or timing remained unclear. 5 means an appropriate ask led to an agreed next step with ownership and timing, or an explicit decline. Do not require a purchase for a strong score.
Use 2 and 4 for behavior between these descriptions. Explain each score with evidence from this call. If evidence is insufficient, say so rather than inventing behavior.
For each criterion, write a short paragraph about what happened and a short paragraph with one practical improvement or behavior to repeat. Include a brief example phrase when helpful, labeled "Suggested wording." Keep advice specific to the call and aim for no more than 80 words per criterion.
Each scored criterion can include coaching in its explanation. If you want additional sections of advice without grades, coaching feedback is currently in preview and must be enabled for your organization. Writing "do not grade this section" inside the scoring instructions alone does not reliably create an unscored section.
5. Compare the results with your own assessment
Save your changes and review newly generated Scorecards for this Call Type. For an existing appointment, regenerate the Scorecard when that option is available; editing the Call Type does not rewrite earlier results automatically.
Start with a strong call, a call with a clear missed opportunity, and one with limited evidence. Before reading the Scorecards, note the behaviors you expect them to recognize. Check whether scores match your standards, explanations cite real moments, and advice gives the rep something they can practice.
If a result misses the mark, revise the relevant criterion and compare again. Keep names and score definitions stable while comparing calls; changing the rubric changes what the scores mean.
After it's done
Use the scores and explanations together when reviewing calls. Clearer instructions help you specify your standards, but the results remain AI judgments: check the transcript before drawing conclusions about someone's performance.
If something goes wrong
- Scores feel arbitrary. Define what earns each level for the specific criterion, rather than just asking Call-E to "be strict."
- Feedback is generic. Request one observed moment and one concrete next action, with suggested wording where useful.
- The Scorecard assumes past behavior. Keep the rubric about this call. It cannot compare against previous Scorecards or client history.
- The feedback is long or repetitive. Reduce overlapping criteria and specify a short explanation with one priority per criterion.