What is Inferred Customer Satisfaction (iCSAT)?

Inferred Customer Satisfaction (iCSAT) is a proprietary, holistic score used by Level AI to measure customer satisfaction by analyzing conversations using generative AI, combining three core signals:

• Customer sentiment (emotional tone)
• Customer effort to resolve the issue
• Issue resolution status

This score is measured on a scale from 1 (very dissatisfied) to 5 (very satisfied). By blending these elements, iCSAT provides a holistic and accurate view of the customer's true experience, overcoming the limitations and biases often found in traditional post-interaction surveys

How Inferred Customer Satisfaction (iCSAT) Is Calculated

The iCSAT score is derived by combining and blending three core signals into a comprehensive measure of the customer experience:

1. Sentiment Score (emotional tone): this detects the emotional tone and shifts in customer feelings throughout a conversation.
2. Customer Effort Score (CES): This measures the amount of effort customers put into resolving their issues. Factors considered include repetitions, hold times, and the number of transfers (a lower effort score indicates a better experience).
3. Resolution Score (issue status): This evaluates whether the customer's issue was successfully addressed, either fully or partially, by analyzing conversation outcomes and final interaction states.

Does iCSAT Analyze All Interactions?

iCSAT is based on the analysis of 100% of customer interactions, as it collects and analyzes unfiltered feedback directly from conversations (calls, chats, emails, etc.) rather than relying on a small, potentially biased subset of responses from traditional post-interaction surveys.

How iCSAT Differs From Traditional CSAT Surveys

Unlike traditional surveys that rely on a small number of post-interaction responses, iCSAT captures real-time sentiment, effort, and resolution signals from every interaction, giving a broader, unbiased view of customer experience.

This approach helps teams uncover root causes of dissatisfaction and track trends without needing to ask customers directly.

How iCSAT Can Be Used to Identify Root Causes of Dissatisfaction

iCSAT breaks each customer interaction into three signals, Sentiment Score, Customer Effort Score, and Resolution Score, and blends them using AI. Because it’s a composite score, teams can look at each component to understand exactly why a customer was unhappy instead of relying on a single survey metric.

A low Sentiment Score points to emotional issues, such as frustration or feeling unheard. A low score on Customer Effort shows the customer had to work too hard, often due to transfers, long hold times, or confusing processes. A low Resolution Score means the problem wasn’t fully solved, which may signal knowledge gaps or system limitations.

By reviewing these signals, teams can pinpoint coaching needs, find broken processes, and uncover unmet customer needs. iCSAT also highlights specific conversations for review, helping leaders understand common situations that reduce satisfaction and take targeted action.