A pilot/copilot model for the data-rich ECC and the mobile command network it may become.
By Mark J. Fletcher, ENP
For several years, I have been circling a question that now feels less theoretical:
As the emergency communications center receives more data, is one call-taker still the right operational unit?
Next Generation 911 is changing what can arrive with an emergency request. Voice is no longer the only source of information. Text, location data, photographs, video, sensor information, language tools, supplemental databases, and other real-time intelligence can all contribute to the operational picture.
That is progress, but every new input creates work.
Someone must receive the information, determine whether it is relevant, validate it, reconcile it with other information, decide who needs it, and communicate it without distracting from the person asking for help.
The information may be useful. It is not free.
We should not respond to this expanding workload by quietly adding every new task to the person who is already speaking with a frightened or injured caller.
The call-taker remains the pilot in command
One possible answer is a pilot/copilot model for the ECC.
The primary call-taker remains focused on the caller. That person manages the conversation, asks the questions, provides instructions, recognizes changes in the caller’s condition, and makes the operational decisions associated with the call.
The primary call-taker is the pilot in command.
The second position functions as a co-dispatcher or data analyst. That person works over the call-taker’s shoulder, either physically or virtually, managing the information surrounding the conversation.
The co-dispatcher might:
- Review location confidence and conflicting location sources.
- Examine incoming photographs, video, or sensor information.
- Compare multiple calls reporting the same incident.
- Identify contradictions between callers.
- Locate specialized resources.
- Monitor changing conditions.
- Validate information produced by automated systems.
- Coordinate with field supervisors or neighboring jurisdictions.
- Present the primary call-taker with concise, actionable findings.
The co-dispatcher does not replace the call-taker or take control away from them. The purpose is to give the primary telecommunicator another trained set of eyes, ears, and hands.
Let the machine do machine work
This is also a natural early role for artificial intelligence in the ECC.
AI can monitor data streams, identify patterns, compare records, transcribe conversations, translate languages, flag inconsistencies, calculate distances, and display possible relationships that a person may not have time to discover manually.
Let the machine do machine work.
But presenting a recommendation is different from taking an action.
An AI system should not independently change a call’s priority, dispatch or cancel a unit, close an event, or convert an unverified allegation into an operational fact. Those decisions require accountable human judgment.
The primary call-taker remains in command. The human co-dispatcher examines the machine’s output, challenges it when necessary, and determines whether it deserves the pilot’s attention.
Simply placing a human somewhere in the process is not enough. That human must have the time, training, authority, and information needed to exercise meaningful judgment.
A copilot does not always mean a one-to-one pairing
It may not be practical or necessary to assign two people permanently to every call-taking position.
A data analyst might support two, three, four, or five call-takers, depending on call volume, incident complexity, and the amount of supplemental information being received.
Two analysts might work together to support a larger call-taking team. During ordinary operations, they could monitor the entire environment. When a significant incident develops, one analyst could become dedicated to that event while the other continues supporting the remaining workload.
The correct ratio would have to be established through actual operational testing. It should not be determined by what fits most conveniently into an existing staffing model or vendor presentation.
The goal is not to create another rigid position. It is to create a scalable human decision layer between expanding data sources and the people responsible for managing emergency calls.
Scott Brillman moved the question into the field
A recent LinkedIn discussion started by Scott Brillman, ENP, CEM, caused me to look at this concept from another direction.
Scott raised concerns about streaming live 911 calls directly to police officers, firefighters, and paramedics while they are responding.
The argument for that capability is understandable. Responders receive information faster and hear the caller firsthand.
But faster access is not automatically safer access.
A responder driving lights and sirens is already processing traffic, navigation, radio communications, CAD information, pedestrians, other responding units, and the physical responsibility of arriving safely. Adding the raw emotion of someone screaming during the worst moment of their life may introduce distraction, emotional priming, confirmation bias, or information overload.
There is also a fundamental information problem. A 911 call represents one caller’s perspective. There may be several callers reporting the same event, each with different and sometimes contradictory information.
Which caller should the responder hear?
One of the critical responsibilities of the ECC is turning multiple information streams into actionable intelligence. Simply bypassing that function and sending raw information directly to the field may move data faster while reducing its operational context.
Scott’s discussion extended my co-dispatcher idea beyond the ECC.
If a second trained person can help manage information inside the center, could that same function sometimes exist in the field?
When does dispatch become mobile?
Imagine a supervisor or command vehicle with a trained dispatcher or data analyst in the passenger seat.
That vehicle might be assigned responsibility for a patrol sector. Calls could be geographically routed to the mobile position, which could coordinate with the primary ECC, field supervisors, responding units, and supporting agencies.
If a major incident developed, the vehicle could immediately become a tactical dispatch position or a mobile command-and-control node.
Instead of asking an officer who is driving to interpret raw caller audio, the trained professional in the second seat could receive and analyze the information, compare it with the larger operational picture, and give the responder what is relevant.
That is a very different proposition from simply putting another audio stream into the driver’s ear.
A mobile dispatcher could potentially provide value during:
- Hurricanes, wildfires, floods, and other natural disasters.
- Large planned events.
- Extended tactical incidents.
- Communications-center evacuations or failures.
- Operations in geographically isolated areas.
- Temporary concentrations of calls in known trouble spots.
- Multiagency incidents requiring local coordination.
- Deployments in which the fixed ECC becomes disconnected from field conditions.
This would not replace the traditional ECC, nor would every patrol vehicle need a dispatcher. The more interesting possibility is a distributed dispatch architecture in which trained people and decision-making functions can move to where they provide the greatest value.
The ECC may become a network
Today’s broadband, wireless, deployable communications, and low-Earth-orbit satellite services are steadily removing the technical barrier of where public-safety work can occur.
That does not mean connectivity alone makes the model safe.
A distributed ECC would still need clear operational ownership, secure access, reliable recording, synchronized CAD records, audit trails, cybersecurity controls, backup communications, defined transfer procedures, and a common operating picture.
The workforce may be distributed. Operational truth may not be.
Everyone involved must know who owns the incident, who is authorized to make each decision, and what happens when connectivity fails. A mobile node cannot become a separate island with its own incomplete version of the event.
The architecture matters more than the furniture. Whether the analyst sits beside the call-taker, supports several positions remotely, works from a regional data pod, or rides in a mobile command vehicle, the fundamental responsibilities remain the same:
Protect the caller relationship. Reduce cognitive overload. Validate information. Preserve human authority. Maintain one operational record.
Pilot the operating model, not the gadget
This concept should not be evaluated through a technology demonstration alone.
A meaningful pilot would need to measure:
- Call-processing quality and time.
- Accuracy of location and incident classification.
- The handling of conflicting reports.
- Cognitive workload for telecommunicators and responders.
- Response-time improvements.
- Driving distraction and responder safety.
- The effect of AI recommendations on human decisions.
- Communications failures and recovery.
- Employee wellness, fatigue, and retention.
- Staffing, training, connectivity, and equipment costs.
- Whether the model produces better outcomes rather than simply more information.
The ratio of analysts to call-takers should be tested. The kinds of incidents receiving copilot support should be defined. The authority of mobile positions should be documented. There should also be stopping rules for situations in which the model creates more risk than value.
A few compelling anecdotes are not enough. Neither is a shorter response interval if it comes with greater driving risk, degraded call-taking, fragmented records, or an unsustainable staffing burden.
Are we finally at the intersection?
I remain enthusiastic about technology.
NG911, AI, broadband, deployable networks, and satellite connectivity can expand what public safety knows and where public-safety work can happen.
But more data is not automatically better information. Faster is not automatically safer. Mobile is not automatically resilient. Access does not automatically equal appropriate use.
The opportunity is to design the human architecture as deliberately as we design the technical architecture.
The primary call-taker remains the pilot in command. The co-dispatcher manages the growing information load, validates machine output, and provides another set of trained hands. AI presents and computes but does not act.
That function might support several call-takers from inside an ECC. It might become dedicated to one high-consequence event. It might operate from a regional support center. In certain circumstances, it might move into the passenger seat of a command vehicle.
Scott Brillman’s discussion helped extend an idea I have considered for years from the ECC into the field. I do not believe every agency needs this model, and I certainly do not believe every call or patrol vehicle requires it.
I do believe we may have reached the point where it is operationally realistic enough to test.
If technology is removing the barriers around where dispatch can physically exist, where should human judgment live?
That is the conversation I hope we are ready to have.