19 July 2023


Let us start with the definition of the word ‘crisis’, according to Thierry Pauchant (PhD, Professor HEC Montréal), a crisis is defined as: 

“An accumulation of probable events at a specific level of an organisation or the organisation as a whole, which may interrupt present or future operations of the enterprise by affecting individuals and communities at a physical, psychological and/or existential level.” ¹ 

According to Thierry, a crisis is a complex dynamic process that is characterised by four phases: incubation, trigger, acute phase and recovery.  

Whether for a company or the society in which we live, crisis management is prepared in advance and is organised as follows: First, it is necessary to set up a team that will be in charge of developing a Crisis Management Plan (CMP) to protect people, infrastructure and assets. The common thread throughout the CMP will be the maintenance of operational conditions. The team will rely on relays identified in each operation and will have actions to implement such as a communication plan. The CMP will be tested several times before being approved. It will be updated according to the criteria defined in advance and will be tested regularly. 

In early 2023, a public organisation in charge of the security of the Olympics Games in Paris asked French data processing specialists such as OPPSCIENCE, to build an analyses and response model in the case of a crisis. The topics covered were a fire at an Olympic site, a terrorist attack, a power outage due to a cyberattack, or a combination of these different cases. The objective was to process all operational documentation. 

The first data source contains the CMP established by the security organisation, but also by all the following entities: Paris Fire Brigade, police, gendarmerie, civil security associations such as civil protection, hospitals, and large public or private sector companies such as Régie Autonome des Transports Parisiens (RATP), Société Nationale des Chemins de Fer Français (SNCF), Orange, EDF, etc.  

Then the following are added: organisational charts, site plans, contingency plans, memoranda, and escalation procedures. All this data is transferred to SpectraOPPSCIENCE’s Intelligence Analysis Management solution ― to be read and analysed semantically in order to link subjects together. This helps to prepare answers to questions asked throughout the crisis, and to ensure that users are not overwhelmed by unnecessary information, hence the need to correlate and prioritise available information

The second source enriches the overview of situations with external data. Spectra repatriates real-time data from news sites, social networks, and official communication channels, all of which are continuously scanned to detect published information that is within the scope of the crisis. Cross-checking information allows users to be informed of other possible crises that took place at the same time.  

The third source is updating the involvement of the mobilised means (material and human) of law enforcement and first responders. Spectra updates in real time in order to share information with the crisis command centre, enabling an overview of the resources deployed and an estimate of available resources and their deployment time. Using this estimate police and first responders can ensure the management of daily missions while dealing with a crisis

A real-life example: When the Notre-Dame de Paris fire struck, nearly 600 Parisian firefighters were at the scene to help extinguish the fire. They were immediately supported by first responders from neighbouring suburbs to help continue daily operations.   

Spectra solution highlights data and the links between the different data. It proposed assessments and scenarios from CMPs to the crisis management team. By cross-checking information, the team was able to consider potential solutions quickly and easily. 

All this is made possible thanks to the solution that is capable of integrating large volumes of raw data from a variety of sources, adapting it so that it corresponds to specific use cases, to proactively provide shortcuts to potential actions and decisions.  

In order for this to happen, it is important to have the right technologies:  

  • A search engine with advanced features allowing the user to quickly access necessary information  
  • A semantic analysis tool able to create a connection between the information in various documentations. This is particularly essential, as today 80% of available data is unstructured (mainly in text format). Semantic analysis extracts rich and/or complex entities, i.e. an entity and its attributes, for example, an emergency vehicle (entity), its crew, its specialisation and its intervention limits (attributes). There can be several types of entities depending on the needs of the users: places, sites, equipment, impacts, etc. Semantic analysis also allows the extraction of older events. The relationship between entities and events are detected, enabling the creation of a crisis-related knowledge base. 

Spectra’s capacity of disambiguation is crucial because it can create a connection between several references without losing the meaning. For example, when a document states that ‘the orange network’s communication antennas are no longer working’ or ‘the criticality colour of a crisis or sub-crisis is orange’, in both cases, the word ‘orange’ is analysed in its context as a communication relay and a critical code. 

  • Machine learning capabilities make it all work independently ― the system does not follow instructions but learns from experience, and its performance improves as the system becomes exposed to more and more data.  

Spectra helps to maintain and increase the operational efficiency of organisations. In a world where available information increases exponentially, our solution assists by quickly extracting essential elements in a data stream. The intelligence provided by this technology ecosystem helps improve the overall quality of the data used to make the right decisions