For decades, computer-aided dispatch (CAD) systems and records management systems (RMS) have been foundational technologies in the public sector (public safety, justice, and government). Telecommunicators depend on CAD systems to capture information, classify incidents, recommend resources, and manage emergency response. Law enforcement officers, court clerks, and jail officials rely on RMS to document incidents and create records that follow cases through the justice ecosystem.
These vital systems almost certainly will look dramatically different in the future.
Artificial intelligence (AI) potentially will reshape both CAD systems and RMS, automating many of the manual processes that burden personnel today. However, the result likely will not simply be better versions of existing systems. Instead, AI is expected to fundamentally change how personnel interact with these platforms while enabling them to focus more of their attention on the people they serve — this will be especially true within the 911 community.
In other words, as agencies explore AI for 911 dispatch and the broader future of 911 technology, the most promising direction is assistance that keeps the telecommunicator in control — human-in-the-loop AI in public safety — not automation that replaces professional judgment.
One way to envision the future of CAD is as an “AI angel on the shoulder” of 911 telecommunicators.
Today, telecommunicators must listen to emergency callers while simultaneously entering information into their CAD system, selecting incident types, determining appropriate classifications, dispatching response, and performing other tasks. In the future, AI-assisted 911 dispatch could perform much of that work in the background.
As a 911 call unfolds, AI could listen to the conversation and make recommendations regarding questions the telecommunicator should ask or flag information that needs clarification. It also could retrieve information that traditionally requires manual searches. Eventually, AI could help determine incident types and recommend appropriate responses based not only on the immediate call but also on other available information.
The telecommunicator would remain the final authority. But instead of concentrating on typing and navigating systems, they could concentrate on what humans do particularly well: communicating with another person during what may be one of the most difficult moments of that caller’s life.
That represents a significant change in how we think about CAD. The technology increasingly would work in the background, supporting the telecommunicator rather than demanding so much of their attention.
That vision won’t — and shouldn’t — happen overnight. AI can make mistakes, produce misinformation, and generate inaccurate results. In a 911 environment, the consequences of those errors could be significant and even life changing. Consequently, the public safety sector — and the 911 community in particular — needs to approach AI adoption carefully.
The answer is not to avoid AI. It is to start with small, deliberate steps.
Agencies should develop strategic technology plans that look well beyond what AI can accomplish today. Leaders should ask where they want their organizations to be in 10 years and identify the steps required to reach that destination.
Part of that process involves training AI using appropriate data, reviewing its performance, and gradually building confidence in its capabilities. Human oversight remains essential. Agencies can introduce AI into specific functions, review what it produces, correct mistakes, and use those lessons to improve the technology over time.
AI already is beginning to enter the 911 environment. Tools can answer nonemergency calls, perform quality assurance (QA) functions, and help evaluate calls. These early applications provide an opportunity for agencies to learn how to use AI responsibly while keeping humans firmly in the loop.
QA provides a good example of how AI could evolve. Traditionally, agencies can review only a portion of the calls handled by telecommunicators. AI quality assurance for 911 calls makes it possible to examine far more interactions. The technology can listen to calls, evaluate them, and generate QA information immediately after the interaction.
Over time, that capability could become real-time assistance.
Instead of simply identifying after the fact that a telecommunicator failed to verify an address, AI might provide a prompt during the call. Or it might suggest that the telecommunicator clarify something the caller said or point out an important detail that could have been missed.
That’s where speech analytics for 911 calls can evolve into real-time call guidance for telecommunicators — and, in some centers, carefully governed AI call coaching for 911 that supports performance without undermining professional discretion.
AI also might identify when a particularly difficult call has affected a telecommunicator. By evaluating factors such as voice inflection, word usage, and the content of a call, AI could alert a supervisor that a telecommunicator might need someone to check on them.
In each case, AI becomes an assistant to, rather than a replacement for, the human professional.
AI also could change how CAD systems leverage information to recommend responses. Today, such recommendations generally depend on information such as incident type and location as entered by the telecommunicator. In the future, AI in computer-aided dispatch (CAD) could consider additional information, including previous incidents associated with an address.
Imagine a call from a location with a history of domestic violence incidents. The initial call might appear relatively routine, but historical information could provide important context. Rather than requiring a telecommunicator to manually search for those records, AI could surface relevant information automatically and in real time.
Ultimately, AI could do much of the “heavy lifting,” i.e., listening to the call, identifying relevant information, retrieving records, and recommending an appropriate response.
That would allow emergency communications centers to make greater use of the skills for which telecommunicators were hired in the first place. Their value would be less about how quickly they can type or whether they navigate software perfectly and more about their ability to communicate, listen, ask the right questions, and obtain critical information from callers.
In practical terms, AI CAD for public safety can show up as 911 CAD system automation and CAD decision support with AI — with the key requirement that recommendations remain explainable, configurable, and easy to override.
The transformation will not stop with CAD. AI in records management systems (RMS) also could reduce the administrative burden on law enforcement officers. Rather than manually constructing a written report, an officer could describe an incident verbally, and then AI could transform that narrative into a professionally written report, automatically placing information into the appropriate RMS fields.
That could include AI police report writing and automated incident report writing that accelerates documentation — while still requiring review and approval before anything becomes part of an official record.
The system also could flag potentially problematic language, including wording that might appear demeaning or demonstrate bias, which could be problematic during a trial or other court proceeding.
Such capabilities could save officers significant time while creating cleaner, more consistent documentation for the court system.
But perhaps the larger and much more impactful opportunity involves connecting RMS with the broader justice ecosystem. CAD, RMS, jail management systems, court systems, mental-health resources, and other platforms often operate independently — in fact, they’re often siloed with no ability to exchange data. Better system integration achieved through effective leveraging of AI will enable sharing of appropriate information across those environments while protecting information that should remain confidential.
That could help authorized decision-makers to see a more complete picture instead of viewing one isolated piece of information. AI also could help connect records that otherwise might remain separated because of issues such as a misspelled name or incorrect date of birth.
This is where CAD and RMS integration, and even AI for justice system data sharing, can start breaking down public safety data silos — but only with careful policy, governance, and security controls.
Achieving this future will require more than better technology. Historically, public safety organizations have kept data in silos partly because of legitimate concerns about confidentiality and security. Technology increasingly may be able to determine which information can be shared and which must remain protected, but such organizations still need leadership and governance structures that enable appropriate information sharing.
Leaders must agree that sharing should happen in the first place. That requires agencies to look beyond immediate technology needs and develop a longer-term vision. Instead of asking only what AI can accomplish today, public sector leaders should consider where they want the technology to take them tomorrow. They also need governance, policies, and safeguards that ensure AI and data are used appropriately.
The human-in-the-loop principle remains critical throughout this evolution. Public sector officials cannot simply “hand over the keys to the kingdom.” Confidence must be earned through experience, testing, quality assurance, and demonstrated performance.
That includes AI governance in public safety, as well as rigorous approaches to public safety data privacy and security — especially when AI touches call audio, CAD notes, and RMS records.
Paradoxically, one of the biggest benefits of greater automation may be making public safety work more human.
For telecommunicators, AI could reduce the typing, searching, classification, and administrative tasks competing for their attention. For law enforcement officers, it could reduce the time required to turn information into professional reports. Across the justice system, it could help people find and connect relevant information that previously remained trapped in separate databases, often with devastating results.
The objective is not to use AI simply for the sake of saying that you’re using the technology, which is analogous to the “shiny object” phenomenon. AI needs to provide real benefits to the public and to the people doing the work.
Getting there will require careful adoption, strong governance, human oversight, and a willingness to envision what public sector operations — especially public safety operations — could look like years from now.
CAD systems and RMS may remain essential components of public safety/public sector operations for years to come. But how people interact with them most likely will change profoundly in the coming years. As next-generation 911 AI takes on more of the manual work happening in the background, public sector professionals could spend much more of their time doing what technology cannot replace: listening, communicating, exercising judgment, and helping people when they need it most.
Jim Lake is an MCP senior consultant. Prior to joining the firm, he spent the last 18 years as director of the Charleston County (S.C.) Consolidated Emergency Communications Center. Email him at JamesLake@MissionCriticalPartners.com.