As they do every year, MCP subject-matter experts attended APCO 2026 in San Antonio, Texas, to see what's new in the public safety technology market. This year's show made one thing clear: the big story is public safety AI, but the real change is the transition towards a platform and integration phase.
The main takeaways are below.
The biggest shift is in how vendors are positioning AI.
Rather than adding AI features to computer-aided dispatch (CAD), 911 call-handling, records management, radio, and other applications, vendors increasingly seem to envision AI as the underlying platform connecting those functions.
Motorola and Axon provided the clearest examples of this trend. Both described environments in which AI spans multiple public safety applications rather than sitting atop individual solutions. The implication is potentially profound: over time, familiar application boundaries — CAD, records management, call-handling, radio, and quality assurance — could become less distinct as capabilities are orchestrated through a common AI layer.
RapidSOS provided another signal. Its CEO demonstrated an experiment in which AI created a rudimentary CAD system, suggesting how dramatically AI-assisted software development could change how public safety applications are built.
Similar themes emerged at NENA 2026, where MCP experts observed growing emphasis on connected platforms, data integration and AI-enabled workflows designed to reduce operational complexity.
Read our practical framework for evaluating public safety AI vendors
Motorola and Axon's portfolios are increasingly overlapping, but their starting positions are quite different. Motorola’s heritage is in land mobile radio (LMR), and it has an extensive portfolio of established public safety applications and customers. Meanwhile, Axon comes from the world of cameras, digital evidence, and newer cloud/AI technologies.
In short, Motorola is bringing its public safety environment toward AI, while Axon is bringing an AI-centric technology model deeper into public safety. Despite their different starting points, both are building increasingly comprehensive platforms, creating greater overlap in the capabilities and customers they serve.
It is no longer the “shiny new object.” Agencies are increasingly asking a fundamental question about AI: What problem does it actually solve?
Earlier AI applications often centered on broadly available capabilities such as transcription or translation. At APCO 2026, it was clear that more vendors are building solutions around specific operational problems rather than simply adding an AI-branded feature. That distinction matters: the emerging question is less about whether a solution contains AI and more about whether AI delivers a measurable operational benefit.
It continues to advance how agencies can review more activity, focus human attention where it matters most, improve coaching, and reclaim significant supervisor time.
AI-enabled QA systems could examine far more 911 calls than human reviewers, flag events that deserve attention, score calls against defined criteria, provide transcripts, synchronize radio traffic and other incident information, and direct supervisors to the specific portion of an interaction that needs review. Some tools also can provide telecommunicators and supervisors with rapid feedback and identify training needs.
The operational benefit is straightforward: an agency unable to manually review even a small percentage of its calls could potentially analyze all of them automatically and devote scarce human attention to exceptions and higher-risk incidents.
NICE's "AI Witness" concept reflects a related need: auditing where AI participated in an incident, what recommendation or decision it made, which vendor's AI made it, and whether the outcome was appropriate. That points toward AI governance becoming an operational capability rather than merely a policy exercise.
Learn more about the key opportunities for AI in public safety from our 2026 MAPS Report.
Another compelling capability involves assembling information from many sources into a coherent incident view.
Demonstrations described searching relatively vague criteria and retrieving the relevant incident along with associated 911 calls, CAD information, radio communications — and potentially video or other data — synchronized in a single interface.
Related technologies are extracting more meaning from communications themselves. Smart Response Technologies (SRT), for example, demonstrated transcription and analysis across radio, telephone, and other communications, with the potential to recognize keywords, tone, emotion, and contextual sounds.
The broader takeaway is that AI's value may increasingly lie in connecting and interpreting information that agencies already possess, rather than simply generating new information.
Despite speculation that broadband will replace land mobile radio, a different conclusion seems to be taking hold: public safety agencies should use LMR and broadband together rather than treat them as mutually exclusive alternatives.
Properly designed LMR systems continue to provide coverage, reliability, redundancy, and operational characteristics that commercial broadband generally cannot guarantee. Commercial networks are built according to business cases rather than public safety coverage requirements, and outages, maintenance practices, and backup power limitations remain concerns.
Broadband, however, is becoming an increasingly valuable complement. Integrated radios can use broadband when LMR coverage becomes weak, or personnel travel outside their jurisdiction. Broadband also enables faster fleet management and programming, GPS updates, and data applications that would be inefficient over narrowband LMR channels.
So, the emerging model is not LMR versus broadband, but rather LMR plus broadband, with each technology handling what it does best.
One of the strongest nontechnology themes at APCO 2026 was leadership development and the loss of institutional knowledge.
APCO appeared to place noticeable emphasis in its educational conference on developing newer public safety leaders. That comes as a generation of executives and practitioners who entered the sector during its major expansion in the 1980s and 1990s approaches retirement.
The concern isn't simply about replacing people. Public safety agencies traditionally have promoted technically accomplished personnel into management roles without necessarily preparing them to manage people or governance. Now, a real need has emerged for structured leadership development and better ways to preserve institutional knowledge before experienced practitioners leave.
Taken together, these observations point to the public safety sector entering a platform-and-integration phase.
AI remains the dominant technology story, but the conversation is becoming more substantive. The important question is no longer whether vendors can claim AI capability. Rather, it's whether AI can eliminate work, connect fragmented information, improve decisions, strengthen QA, govern automated decisions, and ultimately reshape how public safety systems are architected.
At the same time, APCO 2026 showed that transformation will not simply sweep away the existing ecosystem. LMR remains essential even as broadband expands its role; cybersecurity becomes more important as systems become interconnected; established vendors face aggressive new competitors; and the human side — leadership, expertise, and institutional knowledge — remains as consequential as the technology itself.
These observations reinforce themes MCP has seen throughout 2026. At NENA, the emphasis increasingly shifted to connected platforms, data integration, and responsible AI adoption. At APCO, those trends appeared to be accelerating across a broader public safety ecosystem. The implication for agencies is that modernization increasingly requires an enterprise view of technology, data, governance, and operations rather than a series of isolated technology decisions.
David Jones, Chris Kelly, Rich Cagle, Nick Falgiatore, and Sean Scott contributed to this report.