TensorIoT developed a chat and voice bot for a leading financial company to enhance the productivity of handing recurring customer needs.
How can TensorIoT build a bot that can accurately understand the customer’s needs, gather their information, and respond accordingly?
Our chat-bot was able to assist our client in automating a relatively common occurrence in their call center.
For financial companies, consumer fraud protection is a keystone for keeping customers happy. Our client, a multinational consumer credit reporting company, has a call center that provides services for thousands of customers every day. They were looking to incorporate conversational interfaces and voice bot to assist in handling credit report disputes, typically a relatively simple process with only a few potential outcomes, making the process ideal for automation.
TensorIoT built out an automated, natural language chatbot for a common call center, customers wanting to dispute items on their credit report. The solution needed to accurately and completely capture the context of the customer’s actions so that the solution could assist in quickly resolving their issues. Due to the variations in users responses, abbreviations, and other complications, the solution needed to adjust and improve performance for future callers.
The solution allowed customers to call the client’s service line and use the bot to navigate them to dispute a credit item. User gets transferred to the AWS "Connect" call center infrastructure though Amazon Lex. User then enters some basic identifying info so we can locate their credit report. Once we have their report located and loaded, we ask what we can help them with. To enhance our clients performance via synonym recognition, we built a ‘synonym management tool’ to recognize abbreviations, synonyms, etc.