Your team makes decisions in seconds. Understanding those decisions should not take days.“Why did it take so long?”
For a 911 director, that question can set hours of work in motion. A family is questioning an emergency response. A local official wants answers. You need to establish what happened between the first ring and the arrival of help.
The call recording tells you what the caller said. The CAD history shows entries and status changes. Radio traffic captures information passed to responding units. But the explanation may depend on how those pieces fit together: an address corrected mid-call, new information that changed the priority, or a unit that was unavailable when needed.
Until that sequence is clear, you cannot confidently explain the response. You also cannot judge your team’s performance fairly.
And while someone assembles the records, the center keeps working. A supervisor pulled into the review has less time for the people handling today’s calls.
This is a familiar tension in 911 leadership. Directors are accountable for the whole operation, yet the information needed to understand it is often scattered across systems. As new tools enter the center, including AI, that gap can grow.
Closing the gap starts with three connected questions:
What happened? Was it handled correctly? Can we prove every decision? Together, they take a review from understanding the sequence to assessing performance and supporting the findings with evidence.
1. What happened? Follow the information as it arrived
The final record of an emergency can make the situation look clearer than it was.
An address appears in CAD. A priority has been assigned. Units have been dispatched. Looking backward, it is easy to assume those details were available from the beginning.
The call taker may have started with a frightened person who could not give a location. A second caller may have supplied conflicting information. The dispatcher may have received a critical update after units were already en route.
The order matters. So does the distinction between information entered into a system and information actually relayed to the person who needed it.
A useful reconstruction lets a reviewer follow that progression across the call audio, computer-aided dispatch (CAD) event history and radio traffic. Relevant text exchanges, location updates and screen recordings can add context where available. Records need to be associated with the correct event, with timestamps aligned closely enough to establish the sequence and any gaps made visible.
That gives the director a way to distinguish call-processing time from delays involving unit availability or travel. It also helps answer the question beneath the complaint: was there an avoidable delay, and where did it occur?
Sometimes the review will uncover a mistake. Sometimes it will show a call taker patiently resolving conflicting details or a dispatcher recognizing an urgent change and acting quickly.
Both findings matter. Your people need to know that a review can recognize good judgment as readily as it can identify a missed step. Establishing what they knew at the time gives you the foundation for assessing how they handled it.
2. Was it handled correctly? Find the lesson worth acting on
If an updated location was missed, the immediate response might be coaching. But if the same problem appears in calls handled by several experienced employees, another reminder to “watch for updates” is unlikely to be enough. The way information is displayed or handed off deserves scrutiny.
This is where quality assurance becomes useful to leadership. A review should help distinguish an individual training need from a procedure or workflow that is making the job harder.
That distinction is difficult to make from a few calls selected at the end of the month. It is harder still when reviews happen mainly after complaints. You learn a great deal about the calls that attracted attention, but less about the ordinary shifts where habits develop and smaller problems repeat.
The updated APCO/NENA quality assurance and quality improvement standard addresses consistent evaluation, including stronger calibration and oversight processes intended to support fairness. Its 2025 revision also includes guidance on AI and automation.
[1]For a director, the practical question is how to make meaningful review possible with the supervisory time available.
AI-assisted screening can help bring potentially important calls forward and locate portions that warrant a closer listen. But a flagged omission still needs examination. A transcript may miss words. An automated assessment may fail to recognize how a location was confirmed. Some criteria require information beyond the conversation itself.
Supervisors need access to the underlying evidence and the authority to correct the assessment. Employees should understand the criteria and have an opportunity to discuss the review.
The payoff comes in the conversation afterward. Feedback such as “you need to improve location verification” leaves the employee with little to work with. Listening together to the moment the location became uncertain, then discussing the applicable procedure, gives that employee something specific to use on the next call.
The same approach makes recognition more credible. Showing someone exactly how their questioning helped resolve a difficult call carries more weight than a general “good job.”
Several reviews, considered together, can also reveal something the director needs to change. If a recurring problem appears during busy periods, examine concurrent calls, radio workload, staffing levels and handoffs. Those factors do not excuse a missed step, but they may help explain why it keeps happening.
A center-wide average will not necessarily reveal that pattern. Being able to move from a performance trend to the calls behind it gives you a stronger basis for deciding whether the next action should be coaching, a workflow change or a resource discussion.
It can also help you notice demands that call counts alone miss. Two telecommunicators may handle a similar number of calls while having very different shifts. A series of distressing calls may warrant a private check-in even when there is no performance concern.
These findings make the review useful beyond the individual call. But whether you conclude that a procedure needs changing or an employee acted appropriately, you need to preserve the evidence that supports that judgment.
3. Can we prove every decision? Keep the evidence behind the assessment
A director may understand why a call was handled a certain way. Someone reviewing it months later needs to be able to follow that reasoning, too.
The record should establish what information was available, what action followed and the documented basis for significant decisions. Where it leaves a question unanswered, that limitation needs to remain clear.
This becomes more complex as AI contributes to the work. The National Telecommunications and Information Administration has documented uses of AI intended to help manage 911 call demand and support emergency response.
[2] For leaders evaluating these tools, the ability to review their contribution deserves attention alongside their potential benefits.
Consider a live transcript that gets a critical phrase wrong. The call taker hears the caller correctly and disregards the text. If a later review relies only on the transcript, a sound human decision could look like an error.
The record needs to distinguish what the caller said, what the technology presented and what the employee did.
Before introducing a tool, ask the provider to demonstrate how that sequence would be reviewed. Can your center retrieve the output shown at the time? Can it establish when the output appeared and whether an employee changed or overrode it? Can those records be connected to the call and dispatch activity?
Different records answer different questions. A screen recording can show what appeared at the console. Understanding how a recommendation was generated may require additional logs or vendor documentation. Any limits should be understood before the center depends on the tool.
The review process also needs to account for changes over time. If employees repeatedly correct the same kind of output, that is information worth investigating. It may point to a problem with the tool, its configuration or its fit with local procedures.
Preserving this evidence supports all three questions. It helps reconstruct what happened, assess whether the response was appropriate and explain how people and technology contributed to the decisions.
Give the answer, then put it to work
When someone asks why help took so long, a director needs more than a collection of recordings and timestamps. You need an account that explains the sequence and supports a fair assessment of the people and systems involved.
A practical starting point is one recent call that required information from several systems. Apply the three questions: Can you establish what happened? Can you assess whether it was handled correctly? Can you produce the evidence behind the significant decisions?
Where an answer becomes difficult to establish, you have found a gap worth addressing.
That same review should remain useful after the immediate questions have been answered. It can inform coaching, reveal a weak handoff, support a request for resources or show that a new tool needs adjustment.
The public deserves an accurate explanation. Your staff deserve to have their decisions understood in context. And you need enough time left after the review to act on what it taught you.
For more information on incident reconstruction, quality assurance and operational visibility, explore
NiCE’s emergency communications solutions and where they can help.
For more on this topic, I also encourage you to
watch this on-demand webinar Leading 911 with Clarity.
Sources
[1] APCO International, “APCO Announces the Final Approval and Publication of Quality Assurance and Quality Improvement Program Standard,” November 18, 2025. Discusses APCO/NENA ANS 1.107.2-2025. Read the announcement[2] National Telecommunications and Information Administration, “Improving 9-1-1 Operations with Artificial Intelligence,” August 2, 2024. Read the overview and case studies 
