Hybrid QA™ to Chatbot Integration
Connecting Internal Chatbot to Compliance Policies
A Major Global Banks’s compliance managers were flooded with emails regarding basic employee investment compliance questions.
They needed a question answering system that would answer employee investment questions using existing company policies.
Employee questions were not always explicitly addressed by company policies, as some nuance exists. Exact matches connecting to FAQs, while useful, were not comprehensive.
The Lymba Solution
The Lymba Solution uses automatically generated ontologies to process compliance policies. We converted documents via semantics to RDF, stored them to an RDF Triple Store, and integrated our Hybrid QA™ tool into a Chatbot to retrieve answers. Users were able to ask questions such as “Does TPC approval expire? If I obtained TPC approval yesterday, can I execute the trade today?” Hybrid QA™ outperformed Lucene keyword matching and Deep Learning approaches in both accuracy and response time. Lymba increased chatbot performance from 74% to 91% as well as increasing speed.