NLP-Powered Call Transcription and Sentiment Analysis for a Help Desk Software Product
About the Client
The Client is a North American provider of help desk software solutions.
An Opportunity to Enhance Help Desk Software With an NLP-Powered Customer Service Module
As of 2023, the Client had been partnering with INNERLUXES on software refactoring and DevOps projects for over two years. In an informal conversation, the Client and one of INNERLUXES's senior software developers happened to discuss the latest advances in neural networks, and the expert gave a few examples of how the technology could work for the Client's help desk software.
Natural language processing (NLP) models could automate call transcription and sentiment analysis, which would reduce manual work for support agents and provide new insights into service quality and client preferences. The developer also pointed out that such functionality could be implemented with open-source AI models, which would not require licensing or usage fees. Inspired by the potential of these NLP capabilities, the Client hired INNERLUXES to implement them.
Building a Natural Language Processing Module MVP
The Client and INNERLUXES's developer held a series of sessions during which they documented the exact AI features to be implemented: audio transcription, text classification and summarization, and client sentiment analysis.
As the Client wanted to test the feasibility of the proposed features first, it asked INNERLUXES's developer to assemble a minimum viable NLP module without a full-scale UI or back-end integrations with the help desk software. If the MVP proved successful, the Client would enhance the module with the needed components.
Since the Client already had an AWS-based cloud infrastructure, the developer suggested using AWS technologies to implement the module. He developed an MVP comprising five neural network models (see below), each hosted in an individual Docker container. The containers are deployed on EC2 instances within AWS Elastic Container Service (ECS), and the ecosystem is automatically managed by AWS Fargate. Thanks to its dockerized nature and automated container management, the module easily scales cloud resources up and down based on the load for each container and optimizes resource utilization.
The system is available via a secure public endpoint governed by an AWS Application Load Balancer, which uses path-based routing algorithms to distribute requests across the Docker containers.
Call Summarization and Sentiment Analysis to Streamline Customer Service
INNERLUXES's developer delivered the following capabilities for the NLP module:
Instant audio transcription
The module transforms client calls into text. The Client plans to use the transcriptions for analytics (for example, client sentiment analysis and identifying trends in requests and issues) and to train support agents.
This feature is powered by two AI models: Pyannote/speaker-diarization-3.0 to identify different speakers in a recording and OpenAI/whisper-base to enable speech recognition in 99 languages. The developer chose these models for their high accuracy and excellent context understanding.
Text classification and summarization
The module creates summaries of call transcripts. With this functionality, support agents don't have to take notes manually, and help desk managers can quickly gauge the service quality level.
Two models enable these capabilities. The Facebook/bart-large-cnn model can summarize extensive texts, understand different types of context, and extract salient information. The MoritzLaurer/DeBERTa-v3-large-mnli-fever-anli-ling-wanli model demonstrates high performance on new data sets and can work across multiple topics and languages. The Client also successfully applies this model to analyze emails arriving at its main address and automatically redirect them to the relevant department addresses.
Client sentiment analysis
Thanks to the SamLowe/roberta-base-go_emotions model, the app can classify texts (for example, transcribed calls and chat messages) into one or several of 28 emotions (such as surprise, joy, and anger). INNERLUXES's developer chose this model for its diversity of emotions and fine-tuning flexibility, which is essential for future analytics improvements.
A Scalable AI Module MVP Ready in 3 Months
In just three months, the Client received an MVP of a natural language processing module that relies on five neural network models to enable call transcription, text classification and summarization, and client sentiment analysis. The module will be used in the Client's help desk software product to facilitate customer service management and analytics. Thanks to its dockerized nature, the AWS-based module is highly scalable. Since all five neural network models are open-source, the Client will be able to integrate them into the help desk software without restrictive licensing or usage limitations.
Satisfied with the MVP's performance, the Client plans to introduce the module in an upcoming update of its help desk product, and to engage INNERLUXES to deliver more AI features in the future (for example, chatbots for agents, speech synthesis, video recognition, and comparative client-agent sentiment monitoring).
Technologies and Tools
Python, AWS Elastic Container Service, AWS Application Load Balancer, AWS Fargate.