Metrics & Agent Dimensions

Metrics

Avg Overall Score

The Avg Overall Score represents the average quality score across all assessed conversations during the selected period. The score uses a 0 to 100 scale and reflects the combined evaluation of conversation quality, customer handling, resolution effectiveness, behavioural performance, and operational consistency. Higher scores generally indicate stronger customer interactions and more reliable operational execution.

Conversations Assessed

This metric shows the total number of conversations analysed by Isara during the selected period. These conversations are evaluated using Isara’s behavioural, operational, and quality analysis systems.

Avg Effort Level

The Avg Effort Level measures how much effort customers had to expend in order to resolve their issue. The metric uses a scale from 0 to 10. Higher values generally indicate more friction, increased back-and-forth communication, longer or more difficult resolutions, and higher levels of customer frustration. Lower values typically indicate smoother and more efficient customer experiences.

Override Rate

The Override Rate measures how often actions or responses required human intervention or override. A high override rate can indicate poor human or AI agent behaviour, incorrect recommendations, weak automation configuration, insufficient guardrails, or unreliable AI instructions.

Inconsistencies / Conversation

This metric measures how frequently contradictory or inconsistent behaviour occurs within conversations. Examples include contradictory policy explanations, different resolutions for identical situations, inconsistent escalation behaviour, hallucinated rules, or unsupported operational statements. The metric is calculated as the average number of inconsistencies detected per conversation.

Correction Rate

The Correction Rate measures how often human agents needed to correct previous AI responses or actions. Examples include reversing incorrect answers, correcting policy information, fixing pricing or refund decisions, or re-explaining inaccurate information to customers. Higher correction rates usually indicate quality, governance, or grounding issues.

Skill Gap

The Skill Gap measures the difference between the expected capability level and the demonstrated capability level observed during conversations.

Average skill gap

Negative values indicate that the displayed skill level was below expectations. Values closer to zero indicate stronger alignment between expected and demonstrated capability, while larger gaps usually highlight training, workflow, or configuration improvement opportunities.

Unnecessary Customer Effort

This metric measures friction that could likely have been avoided. Examples include repeated requests for information, unnecessary escalations, circular conversations, or asking customers to repeat details that had already been provided earlier in the interaction. The metric is displayed both as a total score and as an average per conversation.

Necessary Customer Effort

Necessary Customer Effort measures effort that was operationally justified or required as part of the support process. Examples include security verification, legitimate troubleshooting steps, clarification questions, or compliance and identity checks.

Customer Repetitions

This metric measures how often customers repeated themselves during conversations. Examples include repeating account information, restating the same issue multiple times, or repeatedly asking the same question. High repetition levels often correlate with poor resolution quality, weak contextual memory, AI misunderstanding, or escalation failures.

Negative Progress

Conversations in which the agent damaged the customer relationship (rudeness, dismissiveness, etc.). Count and rate per assessed conversation.

Negative Relationship

Conversations in which the agent damaged the customer relationship (rudeness, dismissiveness, etc.). Count and rate per assessed conversation.

Agent Dimensions

Knowledge

The Knowledge dimension measures the agent’s understanding of products, policies, procedures, and operational workflows.

Interpersonal

The Interpersonal dimension evaluates communication quality and conversational behaviour. This includes factors such as clarity, professionalism, empathy, tone, and responsiveness throughout the interaction.

Sensitivity

The Sensitivity dimension measures how appropriately the agent handles delicate or emotionally sensitive situations. This includes interactions involving frustrated customers, security concerns, financial disputes, or escalation scenarios.

Resolution

The Resolution dimension measures the agent’s ability to successfully resolve customer issues. This includes delivering correct outcomes, completing required actions, reducing unnecessary follow-up interactions, and guiding customers effectively through troubleshooting processes.

Relationship

The Relationship dimension measures whether the interaction strengthened or weakened customer trust and engagement. This includes building customer confidence, reducing frustration, preserving loyalty, and maintaining positive customer sentiment.

Skill Shown

Skill Shown represents the overall capability demonstrated during conversations. The value is derived from the combined behavioural and operational scoring dimensions.