Use Case: Automatically respond to in-house queries with over 80% accuracy to lower burden on support staff.
When workers at Sompo Japan’s business locations across Japan had questions regarding insurance plans or office procedures, they would search using a text search system created and managed in-house. If the documentation suggested by the search system didn’t help solve the problem, the workers would need to contact support directly.
However, the system wasn’t very accurate at suggesting useful answers, which meant the support team had to spend a lot of time dealing with in-house queries. Sompo Japan was looking for a better way to deal with this problem.
Sompo Japan trained a Text Search Engine in MAGELLAN BLOCKS with three months worth of customer support data, or about 55,000 question/answer entries. Their new text search system soared to close to 80% accuracy, meaning most users were able to find answers without needing to contact the support team directly. Pleased with the results of this initial trial, Sompo Japan are working to expand their system even more.
Publication date: August 2018
- Text Search Engine
Companies are seeing success with BLOCKS.
Predicting advertising returns with AI
Grand Vision Co., Ltd.
- Ad Returns
Optimizing their support center with AI
JCB Co., Ltd.
- Incoming Calls
Building a better home center with AI
- Image Classification
Quick, automated responses to in-house queries
Sompo Japan Insurance
- Natural Language Processing
Powerful, AI-powered marketing support
Predicting returns for new stores with one button click
Creating Sustainable Cities through Waste Collection Optimization
Predicting school entrance exam pass rates in minutes
- Passing Rate
- Average Score
Optimize Construction Related Soil Transport Using the Power of Quantum Computers