Use Case: SumoPPM AI Agent for Conversation Analysis
in a Call Center
​Scenario:
A large telecommunications company has deployed an AI Agent to analyse thousands of conversations between its call centre agents and customers. This agent, called a “Conversational Analyst,” integrates directly with call recordings and transcripts, generating comprehensive, real-time reports on multiple key aspects.
Key Utilities
1. Friendliness and script compliance assessment:
Each call is transcribed and processed by the AI agent, which analyses the tone and words used by the human agent to measure the level of friendliness and empathy. Using natural language processing (NLP) models, the AI agent automatically classifies the tone of the conversation, identifying whether the agent was friendly and professional or impatient and abrupt.
The AI Agent also checks whether the human agent followed the company's approved script. It analyses whether the suggested phrases were used, whether the specified times were met in each section of the call (introduction, resolution, closing), and whether the points were addressed in a structured manner. Any deviation from the script is reported and marked as an area for improvement.
2. Identification of the most requested products and business opportunities:
Through the analysis of thousands of conversations, the AI ​​Agent identifies patterns in product or service requests, generating a ranking of the most requested products. This insight helps the company focus its marketing and personalisation efforts, highlighting the most popular products in future interactions with customers.
The agent also identifies products or services that are frequently requested but not available in the current offering. These “product gaps” are reported to the product development and strategy teams, presenting valuable business opportunities. For example, if many customers are asking for shorter-term data plans, the AI Agent detects this demand and makes suggestions.
3. Assessing problem resolution:
The AI agent measures the level of resolution in each interaction, classifying each call into categories such as 'resolved first time', 'requires follow-up' or 'unresolved'. This analysis is critical for assessing the effectiveness of each agent, as well as identifying recurring issues that could be addressed through product or process improvements.
In addition, if a customer mentions that they have called multiple times with the same issue, the AI Agent detects these patterns and alerts managers to review resolution protocols and develop more effective solutions. The data is presented in a dashboard that allows analysis of the percentage of cases resolved on first contact, one of the key indicators of satisfaction.
4. Detecting cross-selling and up-selling opportunities:
The agent analyses each call for signs of interest in additional products or complementary services. For example, if a customer calling about a technical issue mentions an interest in a new device or an upgraded plan, the AI agent flags the call as a 'cross-sell' or 'up-sell' opportunity.
This data is compiled into reports that highlight business opportunities and suggest cross-selling strategies for sales teams. The company can then implement targeted actions, such as personalised offers or specific promotions.
5. Average resolution time measurement and agent performance analysis:
AI Agent analyses the average duration of each call and assesses whether the resolution time is in line with the company's efficiency standards. It also compares agent performance, highlighting those agents who are resolving issues the fastest and most effectively.
Results are displayed on dashboards where supervisors can monitor metrics such as 'average resolution time', 'first call efficiency' and 'hold time'. Any significant deviations in resolution time are reported and suggested for additional agent training.
6. Summary insights report for strategic decision making:
All data generated by the AI Agent is presented in visual reports with interactive graphs showing patterns and trends. The company gains insights on:
-The most ordered products and those that should be added to the offer.
-The cross-selling opportunities detected in conversations.
-Customer satisfaction levels based on friendliness and script adherence.
-Resolution rates and average response times.
Each month, managers receive a summary report showing both the evolution of the indicators and specific recommendations generated by the AI agent. For example, if the level of friendliness in a region is low, the system can suggest training programmes to improve the quality of interactions in that area.
Translated with DeepL.com (free version)
Benefits for the company:
-
Operational efficiency: Automated assessment allows you to identify agents needing improvement and areas for optimisation without relying on manual reviews.
​
-
Improved customer experience: Identifying and responding to customer needs in real-time improves the quality of interactions and increases satisfaction.
​
-
Data-driven decisions: Insights into products and services help make informed decisions, expand business opportunities and improve processes.
​
-
This AI agent turns conversations into a strategic resource, enabling the company to refine its offerings, optimise training and continuously improve the customer experience.