In the rapidly evolving world of telecommunications, providing exceptional customer support is paramount for leading companies. A prominent telecommunications conglomerate in the Caribbean, with a strong presence in multiple countries and a commitment to innovation, sought to enhance their call center operations by implementing advanced data analytics. This article explores how the integration of Generative AI (GenAI) transformed the customer support experience for this major player in the telecom industry.
The Challenge: Turning Unstructured Data into Valuable Insights
The telecom giant faced a common challenge in their customer support centers: the inability to perform comprehensive statistical analysis on agent-customer interactions beyond basic metadata, such as call duration or response delay. Agents were initially asked to complete forms during live conversations, but this approach proved inconsistent and inefficient, as multitasking during calls affected the accuracy and completeness of the collected data.
To tackle this challenge, a cutting-edge solution was developed using Generative AI to extract valuable insights from unstructured conversation data. This innovative tool processes phone conversation transcripts, identifies relevant information, and generates structured data for analysis. By classifying each interaction into categories and extracting features like intents, flags, assets, and resolutions, the tool significantly improved the consistency and quality of the data available for analytics.
A Comprehensive Solution: AI-Powered Analytics, Orchestration, and Forensics
The AI-powered analytics solution consists of three main components: the analytics component, the orchestration process, and the Forensics tool.
- Analytics Component: This data extraction process handles client-agent interactions in text form, extracting target features from the calls and storing the results in a structured table. The tool segments interactions into the company’s main markets (mobile plans, home connection, TV) to approach each process with a more targeted extraction.
- Orchestration Process: This component manages and troubleshoots the analytics component, scheduling it to run daily on the latest data. It ensures smooth execution and easy monitoring of the entire process. Also, in case of an error or system downtime, it provides easy reprocess of failed process results and backfilling of skipped data.
- Forensics Tool (RAG): This user-friendly tool allows the client to ask questions in natural language about the interactions. An LLM answers these questions by retrieving summaries of relevant interaction data, providing a quick and easy way to explore the overall status or situation of interactions under a certain aspect.
The solution is built using various AWS services, including Step Functions and Glue Jobs for orchestration and data preparation, Lambda Functions for processing tasks, RDS and Glue Data Catalog for storing intermediate results, Athena for the final Analytics Table, and DynamoDB for process tracking.
Empowering Customer Support with AI-Driven Insights
At Strata, we believe that even when we can develop and provide the best tools to solve a problem or issue, it is when the client is able to utilize and manipulate these tools themselves when the best results are achieved, since it is them who interact with their data and their systems on a daily basis.
The implementation of this AI-powered analytics solution provided the telecommunications conglomerate with numerous benefits:
- Structured Data for In-depth Analysis: The analytics tool enabled the client’s Data Analytics team to conduct statistical analysis on interactions between clients and agents in call centers using structured data that is more informative than previous sources.
- Automated Processing Pipeline: The orchestration component automated the processing pipeline, running daily with the latest input data and ensuring efficient, up-to-date analytics.
- Initial Exploration and Overview: The Forensics tool offered a quick and easy way to explore interaction statuses, providing an initial overview of the data on a specific subject.
Additionally, Knowledge Sharing sessions and detailed Runbooks were provided to enable the client to use, execute, and manage all the components of the tool independently.
In conclusion, the integration of Generative AI in the customer support operations of this leading telecommunications company in the Caribbean revolutionized the way they analyze and understand agent-customer interactions. By transforming unstructured conversation data into valuable insights, the AI-powered analytics solution empowers the telecom giant to continuously improve their customer support and maintain a competitive edge in the industry.