If Skadeförsäkringar is the leading property and casualty insurance company in the Nordic region and also to Nordic customers with international operations.
At If there were problems with the custom built insurance system, WayPoint. The biggest problem was lack of scalability and a redesign of the current search application was identified as a possible solution. From a usability perspective the search solution was poorly designed. It had no free text search, forcing users to specify search criteria before searching. Also, users had questioned the relevance of search results. So, the search solution was slowing the system down and needed fixing. Once this was clear, If saw the opportunity not only to solve performance issues but also to create a better search experience. Hence, the scope broadened to include not only technology but also the user perspective.
“In the insurance industry we lack knowledge of the search area and we need support from an external partner within this field. Findwise has delivered with a good team and a good method. Communication and collaboration have been both smooth and flexible throughout the project”
Patrik Kostic - Head of Applications Architecture at If Commercial
The new and improved search solution is based on the open source search engine Apache Solr combined with our customised components. We extract content from WayPoint to a search index, which we make searchable via a separate web site. We’ve designed the interface to include free text search, filters and functionality such as fuzzy search and did you mean. The setup is service oriented which means IF can easily create a variety of interfaces for different business units such as claims or customer service.
Before (database search)
- Full indexing is slow and time consuming, effecting overall performance of WayPoint.
- Indexing causes transactional locks on database.
- Search does note scale well.
- Occasionally time out for advanced or nested queries.
- No support for adding fields dynamically to database search.
After (search solution)
- Fast full indexing that scales.
- Minimal transactional locks on database.
- Incremental updates can run based on connector rules with same speed.
- Search scales for both indexing and query.
- Fast query response time (tested on 30 QPS with <600 ms responses).
- Optimised for executing advanced queries.
- Search engine supports dynamic fields.