The potential opportunities that Artificial Intelligence offers for Emergency Operations Centers have been the focus of numerous studies. As early as 2019, the creation of a social media-based emergency service integrated into existing Emergency Operations Centers was already being considered. EENA, the European Emergency Number Association, has coordinated a series of AI trials and documented the results, evaluating them from the perspective of full-scale implementation.
Currently, AI has a number of limitations, mainly due to the lack of industrialized solutions beyond the POC (Proof of Concept) stage. However, Call Handling represents a unique case. In Operations Centers, AI can bring tangible benefits to call management, according to the processes characteristic of different centers. Examples include: call routing for the European emergency number 112, urgent triage for Emergency Medical Services, non-urgent triage for the harmonized European number 116117, or technical triage for Fire Services.
In call management, peak situations, such as major emergencies (natural disasters, infrastructure accidents, etc.), are highly significant. Overload and stress can lead to errors and inaccuracies. Peak prediction can be addressed by implementing neural networks capable of deciphering complex dynamics that may lead to abnormal increases in call volume, analyzing multiple variables (caller geolocation, call duration, voice sentiment, etc.). Deviations from standard patterns can indicate an approaching peak and suggest appropriate countermeasures.
During an ongoing call, AI can contribute in various ways, making the call more efficient and reducing sources of error. In some cases, AI not only optimizes the call but also triggers the dispatch of emergency services, potentially saving lives. AI contributes in three main ways:
speech-to-Text: Transcribing calls allows operators to quickly capture relevant information while reducing cognitive load, enabling them to focus on critical actions. This is particularly important when dealing with unknown languages (rare idioms or dialects) or stressful situations that impede comprehension. It is no coincidence that EENA’s pilot trials produced varying results depending on the country and context. Quality improvement depends on proper model training tailored to the local context. Once transcription reliability is achieved, the generated text can assist the operator, create automatic call summaries, and, once archived, be used for operator training and predictive analytics. Forecasting phenomena remains one of AI’s major contributions in operations centers;
assisted Interviewing: AI can optimize the flow of questioning. By analyzing the conversation in real time, it can detect keywords and contextual cues to help the operator formulate questions. The sequence of questions is automatically suggested within a defined protocol. In different triage types, intelligent AI support reduces the risk of omissions and accelerates assessments. AI also extracts relevant information to streamline the completion of forms (patient forms, mission forms, etc.) and resource allocation for potential emergency responses;
audio Filtering: AI can reduce background noise and ambient sounds, ensuring clearer communication. Voice isolation, volume normalization, and recognition of critical sounds (screams, impacts, etc.) improve transcription accuracy and extraction of crucial information.
In second-level PSAPs (Public Safety Answering Points), completing patient or mission forms is a critical task in terms of both accuracy and time. For this reason, AI contributions in this area are of great interest to operations centers.
The use of AI is a defined path, not just one of many technological options. Its potential is enormous, and effective implementation today requires integrated solutions within CAD (Computer-Aided Dispatch) platforms, rather than isolated technologies, which may be counterproductive in relation to the overall process. This entails establishing the correct balance between human decision-making and machine decision-making. The “Human in the Loop” is essential for ensuring effective interventions: AI provides data and suggests solutions, but decisions must remain in human hands.