Most of the companies have customer care departments, which connects with customers for variety of reasons like support, billing, new customer acquisitions. They connect using chat, phone calls, email, mobile apps etc.
Many companies are now trying to utilize the power of Big Data by storing chat logs, emails, audio and call transcripts etc in Hadoop to gain better insights, improve customer experience and improve operational efficiency.
However processing large volume of chat and call transcripts in batch or realtime to determine context and customer sentiment is hard.
InsightLake Sentiment Analysis solution enables customer care departments to process chat and call transcripts in both real time and batch mode. It provides following functionalities:
Sentiment Analysis solution allows operations team to create quick models to understand and extract text from unstructured data sets like chat logs, call transcripts, emails which can be in JSON, XML, Plain Text form. It allows creation
of chained interactions with sub text.
It also allows evaluation of complete text level as well at sub text/interaction level.
Sentiment Analysis solution integrates with SOLR and Elastic Search data stores and stores enriched interactions with text, metadata and generated sentiment so operations teams can do following:
Sentiment Analysis solution allows processing of unstructured email, chat/call transcripts data through Kafka channel.
It allows applying models on unstructured events, enrichment with master data and finally extraction of context, keywords, creation of sentiment scores and determination of high sensitive events.
In case of high sensitive events, actions can be provisioned to create Critical events and incidence for operations team, alerts are posted using emails so agent's live chat and call will get another assistance or intervention from ops team.