clients / case studies / businesses
Kinnevik Group
Goal
Create a flexible, easy-to-use, automated online intelligence gathering systemSolution
Delivered a quality intelligence gathering tool to the Chairman of group and his staffIntroduction
The Chairman of Kinnevik Group wanted to create an intuitive, automated, intelligence-gathering application to effectively automate various research tasks online. A self-made multi-billionaire, the Chairman was very demanding and exacting of the intelligence the tool he desired for his set of business analysis. He needed a tool to continuously search, collect, pre-interpret and summarize the knowledge it would find. His reviews of his initial vendor showed they were struggling in their attempt. He commissioned us to become involved, and upgrade or replace their prototype attempts. It was vital that our solution was on time, within budget, and effective, unlike the previous third party development attempts had been.The previous third party product was unfortunately plagued with reliability and effectiveness issues, and key functionality was not implemented. Ephibian became tasked with fixing a fast and effective tool to automate intelligence gathering for the Chairman and his direct staff.
What did we do?
- Built a fast, effective, flexible, and intuitive Java-based Collection application, which uses a sophisticated threaded queuing mechanism for downloading the pages of searched material with a flexible Keyword search to simultaneously check multiple data sources
- Unlike a single-time search engine, the system runs continously: spidering key sites and news sources constantly for desired materials --- and interpeting the relevancy of results as they are reported
- The system is configurable to maximize the available server hardware and bandwidth, without overloading resources; and allows the system to be scaled in terms of resources-used during offpeak hours
- We enhanced the system’s administrative capabilities
- The System is able to search Web pages (main frame and subframes), Message Boards, leverages search forms, and implements a WebCrawler functionality to search throughout any of its sources
How did we do it?
- Within a number of simultaneous threads is a configurable parameter, which can be optimized for the server hardware and available network bandwidth
- The relevance-matching logic is separated from the Collection logic, and provides summaries, details and confidence factors
- System incorporates an Administrative GUI, configuration utility, error log, and comprehensive statistics log
