Case Study

Improving Telecom Customer Support with Gen AI

The competitive landscape across wireless network operators is increasingly intense. This increased competition has created a commensurate increase in wireless customer expectations. Customers now demand greater speed, higher levels of network reliability, fewer service disruptions, and faster issue resolution. The full optimization of all the above is now required to maintain customer loyalty and grow market share. This rapidly evolving environment requires continuous development of innovative technologies and solutions.

Leveraging Generative AI to provide faster resolution of wireless network performance issues.

Current Challenge

Common Problems Faced by Enterprises

Modern wireless networks deliver services at reliability levels that far exceed their predecessors. However, service issues are still a fact within these complex networks…and customers are increasingly less tolerant of them. Wireless operators face challenges related to….

01

Service Ticket Complexity

Complex networks breed complex issues that require more knowledge and expertise.

02

Scattered Information and Knowledge

Siloing of critical data and historical knowledge inhibits the rapid use of that information to resolve issues.

03

Manual Information Search and Retrieval

Continued reliance on non-automated processes for data search and retrieval creates inefficiencies and error.

04

Over Reliance on Human Subject Matter Experts (SME)

Human intervention too often required to provide insights that would be available via AI-enabled protocols.

All the above create lags in issue resolution times and degraded customer experience.
KMS GenAI
Plasma delivers a Gen AI Knowledge Management Solution (KMS) that has been developed based on a 20-year history of developing high-tech solutions for the wireless network industry. A Gen AI KMS that combines advancements in artificial intelligence, machine learning, and natural language processing to deliver a comprehensive solution that meets the complex requirements of modern wireless networks.

The Solution

Generative AI Knowledge Management

More Robust Issue Resolution - Key capabilities of Plasma’s Gen AI KMS

Accurate and precise access to historical information (tickets, documents, workflows, mitigation information) via retrieval augmented generation (RAG) that leverages existing LLMs (Large Language Models). A RAG process that includes the following:
  • Loading: Getting data and metadata from where it lives – text files, PDFs, other websites, a database, or an API – into the pipeline.
  • Indexing: Creating a data structure via vector embedding (numerical representations of the meaning of data) and other metadata strategies to support quick retrieval of contextually relevant data.
  • Storing: AI-enabled data storage that eliminates the necessity for re-indexing.
  • Querying: Utilization of LLMs and an index to query data (including sub-queries, multi-step queries, and hybrids).
  • Evaluation: Active RAG model evaluation and dynamic fine-tuning to obtain optimum results.
Ability to query information based on AI search capabilities than actively learn and evolve with the expanding knowledge ecosystem.
  • Extract Features from data – text, tables, and images
  • Embed images, text, and tables together via a multi-modal embedder
  • Retrieve applicable images through a “similarity” search
  • Automated mining of unstructured data for hidden insights
  • Provision of customer service that is faster and more precise
  • Ability to conduct customer sentiment analysis on a large scale
All the above combine to provide issue resolution that is less dependent on human intervention.

Overview of Gen AI KMS Workflow

KMS Workflow

The Impact

Impact on Wireless Network Issue Resolution