Implemented a chatbot to fetch data
- The client wanted to overcome the limitations of their current infrastructure.
- Wanted on-demand insights using an interface that is not technically daunting for customer service representatives.
- Needed a way to make searching for data easier even for those with limited knowledge of how set queries into a database.
- Developed a chatbot using IBM Watson development framework and integrated it with their web.
- Used IBM Watson’s internal NLP algorithm and parser to identify the intent and entities in the entered text input in the interface that customer care representatives use to identify the issue.
- Constructed HiveQL queries from the NLP algorithm output.
- Used Webhook to connect to the database and retrieve query results.
- Formatted retrieved data that would be presented to the customer care executive.
Tools & Technologies
- Customer care executives can fetch a specific set of insights by interacting with the bot in plain English.
- Insights available on-demand leading to better analysis and quicker resolution.
- Improved customer experience and better utilization of resources.
- Easy integration with existing infrastructure leading to no extra cost or process changes.