Context Intelligence for Crisis Monitoring and Risk Mitigation

January 21, 2014

To illustrate the variety of application of Context Intelligence we will publish in the coming weeks various case studies.

Our first case demonstrates how Context Intelligence uncovers crisis events, monitors and manages them, ultimately reducing risks.

Company: A large multi-national corporation providing power and energy solutions . The company is operating worldwide with activities also in politically fraught geographies. The company is unfortunately regularly exposed to critical situations in these areas. The corporate communications team is tasked tracking these situations around the globe. For this, the team scans multiple internal and external channels such as press and social media in multiple countries and languages.

Challenge: Current solutions ranging from media monitoring to social media monitoring fail at detecting crises events. While they keep track of what is been said in general terms about the company they disappoint when used to identify key events as they happen and follow key stakeholders. Instead they return too many so called false positives – results seemingly matching the criteria but actually are just noise.

Solution: In the event of a crisis it is critical for the team to track all involved stakeholders and key players such as local subsidiaries including employees, employee relatives, journalists, bloggers, etc. The core request is for a platform, which allows the team in the event of a crisis to quickly identify and monitor key players and key developments with limited manpower. Squirro is such a platform.

Context generation: In the crisis monitoring application a number of internal structured sources, especially CRM data, are used to create a frame of reference. A number of external data sources ranging from newswires, to social media, general web search, specialised community sites and in some countries also chat data are continuously mapped into this information space. The mapping rests on enriched data streams drawing on a number of techniques from data comprehension and enrichment, entity detection, sentiment analysis, natural language processing, and pattern classification. The image below shows the interface trending information on a developing event.


Results: The variable setup allows the company in the event of a crisis to quickly setup a crisis specific monitoring and leverage this across the organisation. The solution allows the company to include all type of external news sources ranging from traditional newswires, to social media and in some countries even the inclusion of chat services. The system in place lets the company include the local subsidiaries and other stakeholders to contribute to this monitoring approach. A federated system setup guarantees local relevance and availability while allowing corporate headquarters to stay in control as the situation evolves. The company realises substantial risk mitigation by staying on top of developments.