Key Terminology
Verbatim
The customer’s stated reason for calling, captured in their own words rather than internal business terminology.
Interaction Classification
OI classifies customer interactions by analyzing their stated reasons for contact at scale using language algorithms. This enables organizations to gain deep insights into customer needs and operational efficiency.
Interactions are classified into three hierarchical levels:
Root Cause
The fundamental reason a customer reaches out—not necessarily the underlying issue.
Think of this as the opposite of the interaction: if this didn’t happen, the customer wouldn’t need to contact support.
Inquiry
The main topic of the interaction, directly linked to the Root Cause.
(For organizations in Australia, New Zealand, or the UK, this may be spelled "Enquiry.")
Sub-Inquiry
A more detailed breakdown or most granular level of the Inquiry.
(In Australia, New Zealand, or the UK, this may be spelled "Sub-Enquiry.")
Example Classification:
A customer calls and says:
"Hi, I placed an order online about ten minutes ago, but I didn’t receive a confirmation. I’m calling to check if my order went through."
This interaction could be classified as follows:
Root Cause: My order was placed
Inquiry: No order confirmation was received
Sub-Inquiry: Has my order gone through
Additionally, based on the call transcript, Predicted Satisfaction and Predicted Resolution scores can be generated.
Predicted Metrics:
Predicted Satisfaction
A score from 1 to 5, where 1 indicates low satisfaction and 5 indicates high satisfaction.
This score is determined by analyzing the language customers use during an interaction.
Unlike surveys (which typically have a <10% response rate), this method allows organizations to predict customer satisfaction at scale while minimizing survey bias.
Predicted Resolution
A binary score of 1 (resolved) or 0 (not resolved), averaged across interactions to measure resolution rates.
How it's predicted:
If a customer mentions needing further action (e.g., calling back, messaging again, submitting a document, or visiting a store), the interaction is marked as not resolved.
Otherwise, it is considered resolved.
Important Note: Unresolved interactions don’t always indicate agent error—they could result from business processes or regulatory constraints.
Agent Effectiveness Score
Measures how efficiently an agent resolves an issue while maintaining high customer satisfaction.
Helps compare agent performance and identify areas for process improvement
A lower score is better in this metric
Agent Effectiveness score is calculated by:
[Handle time / (Predicted Satisfaction + Predicted Resolution)]
Example Comparison:
Call | Handle Time | Predicted Satisfaction | Predicted Resolution | Agent Effectiveness Score |
Call 1 | 600 sec (10 min) | 5 (high) | 1 (resolved) | 100 |
Call 2 | 600 sec (10 min) | 1 (low) | 0 (not resolved) | 600 |
Although both calls take the same time, Call 1 has a better (lower) Agent Effectiveness Score because it resulted in higher predicted satisfaction and resolution.
Sentiment Analysis
Determines the overall sentiment of a customer interaction based on customer language.
Scoring scale:
-1 = Extremely negative sentiment
0 = Neutral sentiment
+1 = Extremely positive sentiment
Opportunity Types
Opportunity Types help categorize Inquiries based on:
Business Value: Does the inquiry benefit the company?
Customer Value: Does the inquiry provide value to the customer?
Understanding these categories allows businesses to identify areas for improvement and measure the impact of initiatives.
Common Opportunity Types:
1. Enable Self-Service
Valuable to customers, but not beneficial for the business to handle via a contact center.
Goal: Shift these interactions to self-service or automation.
2. Make It Easier
Valuable to the business, but time-consuming for customers.
Goal: Simplify these interactions to improve customer experience.
3. Pain Point
Not valuable to customers or the business.
Goal: Eliminate as many of these interactions as possible.
4. Value Generating
Valuable for both customers and the business (e.g., revenue-generating interactions).
Goal: Optimize these interactions to maximize business impact.
Additional Features
Exporting Data
To download data, click "Export CSV".
For more details on exporting chart data, refer to the [Exporting Data Guide].