SmartBot is an intelligent, spontaneous answering machine for the enterprise.
Features: SmartBot’s administration application is built to make the bot easy to train, take corrective measures and analyse the customer behaviour. Admin application has the following features
- Define the meta-data for the business use case. Here the administrator will define various types of conversations possible and corresponding lexicons to each conversation type
- Manage the historical data by uploading the conversations between support staff and customers. This data could be just sample data or real chat transcripts. The tagging of each conversation is done by an intelligent text search engine based on the meta-data. Administrator can correct any wrongly tagged conversations.
- Dashboard with key indicators on the performance of the SmartBot – E.g. Count of happy and unhappy conversations based on the conversation sentiment, geographic distribution of the customer base etc. Administrator can also view all conversations between support staff and customer
Technology: To build the training data, there is a feedback/learning mechanism built on a combination of machine learning and Elastic search engine.
Integration layer is built with REST API to meet the industry standard requirements. There will be a custom development effort to integrate with third party applications.