Call Scoring

User Level: Admin, Supervisor, Analyst

---

Call Scoring automatically evaluates each eligible call and provides fast, consistent feedback. The system uses a company-defined template to score key parts of a conversation (for example: Discovery & Needs Analysis, Value Proposition & Solution Fit, Objection Handling, etc.), adds short written explanations, and attaches transcript highlights as evidence. This helps supervisors focus on calls that need a human review, while agents get quick coaching moments after every call.

​

Prerequisites

• Admin access to AI Conversation Intelligence
• At least one scoring template set up
• CI Plus for scoring template auto-assignment per group (calls handled by human agents). Auto-assignment per Voice Agent is included with AI Voice Agents and needs no CI plan.

​

Setup

1. Go to Settings > Conversation Intelligence 2. Toggle on Call scoring. Under Scoring templates, click New scoring template, or Edit on an existing template. The list shows each template with its number of questions, a Default badge on the default template, and its assignments. The counter shows how many of the 10 templates are in use. 3. Inside the template editor:

• The editor has two parts: Settings (template name, Apply to groups, Apply to Voice Agents, Set as default template for unmatched calls) and Questions (the list of questions, each labelled with its category; + adds a question).
• Categories are fixed (Opening & Framing, Discovery & Needs Analysis, Value Proposition & Solution Fit, Objection Handling, Closing & Next Steps, Tone & Empathy).
• Inside each category, you can add, remove, or edit questions.
• To add: choose from the predefined question library, then edit text/details.

4. Each question is answered on a 5-point scale:

• 5-point scale (Very Poor → Very Good)
  • Define what each level means.
  • Provide at least one example for Very Poor (1) and Very Good (5).
  • Mid-levels (Poor, Fair, Good) can also have examples (recommended).
  • Example:
    • Very Poor (1): “Agent did not explore the customer’s needs at all. Example: ‘Let me show you what we offer.’”
    • Very Good (5): “Agent thoroughly explored the customer’s situation with follow-up questions. Example: ‘What’s your current process like? What challenges are you facing?’”

5.Save the template. The default template applies automatically to any call no other template covers.

​

Applying templates

Up to 10 scoring templates per account. A template is applied to a call in one of three ways:

• Default: one template is the account default, used for any call not routed through a group or handled by a Voice Agent covered by another template. Set it in the template’s Settings with Set as default template for unmatched calls. Only one template can be the default at a time. New accounts start with the Default CloudTalk scoring template (10 questions).
• Manually: a template can be applied to a call from Call Details.
• Auto-assignment: open the template and, under Settings, use Apply to groups and Apply to Voice Agents to choose whose calls it should score. A group or Voice Agent can only belong to one template at a time. Every call is matched automatically to the template of the group or Voice Agent that handled it. Anything unmatched falls back to the default. On CI Basic, a template with a saved group assignment shows a Requires CI Plus badge and those calls are scored with the default template instead.

​

Show call scores to agents

A toggle in Call Scoring settings. Off by default, applied account-wide (not per agent).

• On: an agent with CI enabled sees the score in Call details for calls they took part in. An agent only ever sees scores for their own calls, never another agent’s score, and never the scoring templates.
• Off: the score is not shown to agents.
• The other CI data in Call details (transcript, summary, smart notes, topics, talk/listen ratio, sentiment) stays visible to agents whenever CI is enabled for their user, regardless of this setting.

​

Where to find it in CloudTalk

1. Go to Analytics > AI Analytics.
2. In the list, each call shows a Score badge along with Talk/Listen ratio, Sentiment, Topics, and more.
   
3. Click the three dots on the right of any row to open Call Details, then select the Call Score tab to view the full breakdown. Inside the Call score tab you’ll see:
   • Section scores with labels (Opening & Framing, Objection Handling).
   • A short AI-written summary for each question explaining why the score was given.
   • Transcript extracts linked directly to the relevant moment in the call (click “Check transcription” to jump straight there).
   • The overall score at the top, based on your template.
   

In the desktop Phone app, a call’s AI insights tab also shows a Call Score card; agents see it only when Show call scores to agents is on.

​

How scoring works

​

Template-driven

Each score is based on a scoring template. The template that scores a call is the one that covers it: your default template, a template applied manually from Call Details, or the template auto-assigned to the group or Voice Agent that handled the call.

• A template is made up of fixed sections (for example Opening & Framing, Discovery & Needs Analysis, Value Proposition & Solution Fit, Objection Handling, Closing & Next Steps, Tone & Empathy).
• Inside each section, you add one or more questions.
• Every question is answered on a 5-point scale, from Very Poor to Very Good. For each question you define what each level means and add examples, so the AI scores consistently from call to call.

​

What happens after a call

1. The transcript is generated.
2. The AI applies the scoring template that covers the call.
3. For each question, it checks the transcript for evidence, maps what it finds to the 5-point scale, and writes a short justification with a transcript highlight.

​

Weights and calculation

1. Equal weight for sections. Each section contributes the same share of the overall score. A template with 10 sections gives each section 10%; a template with 5 sections gives each 20%.
2. Equal weight for questions within a section. One question carries the full section weight. Two questions carry half each, and so on.
3. Section score. The average of its questions, rounded to the nearest whole number.
   • Section 1 (2 questions): (25% + 100%) ÷ 2 = 62.5%, rounded to 63.
   • Section 2 (2 questions): (100% + 0%) ÷ 2 = 50%.
4. Overall score. The average of all rounded section scores: (63% + 50%) ÷ 2 = 57%.

​

Answer mapping

Each level of the 5-point scale maps to a percentage:

• Very Poor = 0%
• Poor = 25%
• Fair = 50%
• Good = 75%
• Very Good = 100%

​

Handling Not applicable (N/A)

The AI returns N/A when it cannot score a question (for example, there is no transcript evidence for it) or when a section has no questions. N/A items are excluded from the calculation, and the weights for the section and the overall score are recalculated using only the applicable questions.

​

What CloudTalk generates

For every scored call you get:

• An overall score from 0 to 100.
• A score for each section.
• A short explanation of why each score was given.
• Transcript highlights linked to the exact moment in the call, so you can check the evidence yourself.

​

Best practices for template design

• Make behaviors observable
  • Don’t use vague terms like “Great at discovery.”
  • Do use measurable actions: “Agent confirms the customer’s issue in one sentence.”
• Avoid context-dependent questions
  • The AI only uses the call transcript.
  • Don’t ask: “Did the agent follow the script?”
• Keep scope sharp
  • Fewer, sharper sections (4–7) are better than many broad ones.
• Define the ends clearly
  • Always provide a definition and at least one example for Very Poor (1) and Very Good (5).
  • Example (Objection Handling):
    • Very Poor: “Agent dismissed the objection without acknowledgment.”
    • Very Good: “Agent acknowledged the concern empathetically and reframed it positively.”
• Use multiple examples
  • Especially for Very Poor and Very Good, provide 2–3 examples to reduce ambiguity.
• Plan for N/A
  • N/A is system-handled.
  • Use the description field to clarify when it’s expected (e.g., “Intro not needed on an internal transfer call”).
• Review results regularly
  • After two weeks, check if your template is driving the coaching outcomes you want.
  • Refine vague questions or add better examples if needed.

​

FAQ

• Does the AI score affect call routing or KPIs automatically?
No. Scores are for coaching and analytics.
• When does the score appear?
Call Score is calculated together with other Conversation Intelligence metrics, after the call has ended and once the transcript is available

• Which roles can see scores?
Admins, Supervisors and Analysts can see scoring data in the Dashboard. Agents see the score for their own calls only, and only if an Admin has switched on Show call scores to agents.
• If scorecard questions change after scoring has been applied, will calls be rescored? No. Changes to the scorecard apply only to future calls. Calls scored before the change are not recalculated.

• Can I create custom questions? Yes. You can add a question from the predefined library and then customize it: edit the question text and update the descriptions for each level of the 5-point scale. The only part that cannot be customized is the category, which is fixed.
• Can I delete a scoring template?
  No. To stop using one, remove its group and Voice Agent assignments; the calls it covered then use the default template.

---

Need help or have a question? Just reach out through our Support portal — we’re here for you.