New Orleans Will Use Artificial Intelligence to Handle 311 Public Service Calls
New Orleans is about to change the way residents ask for help with city public services, and artificial intelligence is the star of this transformation.
In just a few months, anyone who calls 311, the city’s non-emergency services hotline, might be greeted not by a human but by an AI agent trained on three years of real call data. The system was designed to be polite, informative, and focused on serving the public. And here’s a fun detail: it knows how to correctly pronounce Tchoupitoulas, one of the most notoriously tricky street names in New Orleans 😄
The person behind the project is Karl Fasold, executive director of the Orleans Parish Communications District, who has a background in computer science and served as the district’s chief technology officer before stepping into his current role in 2023.
The data driving this decision is pretty straightforward: half of all 311 calls are simply requests for information. Nothing that requires a human agent to resolve. According to Fasold, the AI agent won’t generate service request tickets but will provide information quickly and accurately, covering exactly that 50% slice of calls that are purely informational.
But the story goes well beyond 311.
Since 2023, the city has already been using AI to triage 911 emergency calls, and the results are reshaping the daily routine of a team operating with roughly 40% fewer staff than needed, due to fiscal uncertainties affecting headcount.
It’s real efficiency, but it also raises serious questions about transparency, accountability, and the limits of automation when public safety is on the line.
Below, you’ll learn how all of this works, what’s already in place, and why the New Orleans case could become a model — or a cautionary tale — for the rest of the world. 🌎
How AI Is Transforming Public Service in New Orleans
The 311 call system in New Orleans works as a centralized hub for non-emergency public services. Residents call in to report potholes, request extra trash pickups, ask about municipal service hours, or get answers about city hall processes. For years, this massive volume of calls was managed exclusively by human teams who had to deal with repetitive, routine questions that could, in practice, be answered by any well-trained system. That’s exactly where the city spotted a real opportunity to deploy artificial intelligence strategically, with an immediate impact on residents’ daily lives.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
The AI agent that will handle these calls wasn’t built from scratch based on guesswork. It was trained on three years of actual call recordings, meaning the system learned not just the content of the answers but also the context, tone, and nuances of interactions between agents and residents. This kind of training based on real historical data is one of the strongest foundations for ensuring that artificial intelligence delivers useful, accurate, and culturally appropriate responses to callers.
Beyond technical accuracy, the city also paid attention to the experience of the person on the other end of the line. The agent was configured to sound — in Fasold’s own words — customer-service focused, polite, and patient. Those are qualities that aren’t always guaranteed in human interactions under high-demand pressure. With half of all calls being simple information requests, the AI can handle those situations quickly and efficiently, freeing up human agents to tackle cases that truly require judgment, empathy, and real-time decision-making.
The 911 System Has Been Using AI Since 2023 — And the Impact Is Real
While the 311 project is still ramping up, the use of artificial intelligence in 911 emergency calls in New Orleans has been a reality since 2023. And beyond that, the city was one of the first in the United States to adopt AI-assisted emergency call triage.
The technology was developed in partnership with Carbyne, a company that builds software for emergency call centers. Carbyne had already upgraded New Orleans’ legacy 911 system to a cloud-native product and was working on an AI agent when Fasold expressed interest in the concept.
The problem that drove adoption was specific and recurring. Alex Dizengof, co-founder and CTO of Carbyne, explained that when they analyzed the data, they identified a repeated pattern: every morning and every evening during rush hour, a traffic accident would occur and hundreds of people would call 911 to report the exact same thing.
The result was a long queue that buried critical emergency calls.
As Dizengof described it, if someone called in reporting a heart attack, that person went to the back of the line. The queue is blind. Dispatchers had no way to prioritize the calls.
How AI Triage Works on 911
The solution was clever and technically elegant. Carbyne’s system was integrated with the computer-aided dispatch system (CAD) used by first responders. This means the AI agents know, in real time, where every active traffic accident in New Orleans is at any given moment.
Here’s how it works: anyone who calls 911 from within a 200-meter radius of an active traffic accident is immediately routed to the AI agent. The AI asks if the caller is reporting that specific accident. If the answer is yes, the agent lets them know the situation is already being handled and help is on the way. If the answer is no, the call is immediately transferred to a human dispatcher.
Fasold says the impact has been significant. The estimate is that the system has given back at least two hours per day of dispatcher time, which, for an agency running with nearly 40% of its staff positions unfilled, makes a massive difference. Every second counts — both for the overworked team and for anyone calling in a life-or-death situation.
Dizengof highlighted an important positive side effect of this automation: humans get tired of repetitive, mundane work. And when that repetitive work is also critical, fatigue can become dangerous. By removing that burden from dispatchers, the AI doesn’t just improve operational efficiency — it potentially helps save lives by keeping human professionals more alert and available for the moments that truly demand their attention.
The 911 AI Agent Has Clearly Defined Limits
One point Fasold makes very clear is that the 911 AI agent only handles calls related to traffic accidents. It will never be used to answer actual emergency calls. This deliberate limitation is essential for maintaining trust in the system and ensuring that high-risk situations continue to be managed entirely by trained human professionals.
The Bourbon Street Terrorist Attack Incident
One of the most revealing moments about how the system works happened during the Bourbon Street terrorist attack. According to Fasold, the AI agent answered the first two 911 calls before routing them to a human dispatcher. For approximately 20 seconds, the incident was classified as a hit-and-run until human triage identified the true severity of the situation.
This episode raises important questions. Twenty seconds might seem like nothing in a normal context, but in an emergency with victims, every moment matters. At the same time, the system did exactly what it was designed to do: it answered a call within the radius of what appeared to be a traffic incident and, when the situation turned out to be something different, transferred it to a human. The question isn’t whether the AI failed — it’s whether the system’s design is robust enough for atypical, extreme scenarios.
Transparency and Regulation: A Conversation That Still Needs to Happen
Two computer science professors at Tulane University, Aron Culotta and Nick Mattei, who also co-direct the university’s Center for Community-Engaged Artificial Intelligence, offered a balanced perspective on the topic. They noted that information technology has been used to reduce human contact since call centers gained traction in the 1980s, and that many people today are perfectly comfortable installing Ring cameras or interacting with assistants like Alexa.
Mattei pointed out that the discomfort in this case might stem from the fact that it’s the government doing it — not something you bought and brought into your home by choice. But for very narrowly defined applications, he believes this could be a powerful way to free up government resources. Just because it’s AI doesn’t mean it’s bad, he added.
However, there’s a sensitive point that can’t be ignored. Currently, anyone who calls 911 or 311 in New Orleans is not explicitly told when they’re speaking with an AI system. While the voice is obviously non-human, there’s no formal notification. On top of that, the communications district has no plans to publicly announce when the 311 AI agents go live.
There’s also no specific municipal policy on the use of AI in emergency services, and national regulation on the subject is still virtually nonexistent.
Culotta emphasized that the responsibility falls on the public to demand transparency and accountability. It’s a stance that places the weight of oversight on citizens — something that raises its own questions about the obligation of government to be proactive in communicating how its operations work.
New Orleans as a Global Reference for AI in Public Services
What’s happening in New Orleans isn’t just a local technology modernization initiative. It’s a real-world experiment that other cities around the globe are watching closely, because the challenges faced there are universal. Staffing shortages in public service call centers, growing call volumes, the need for faster responses, and pressure for operational efficiency are problems that affect municipalities of every size, on every continent.
The choice to start with 311, the non-emergency services channel, was strategic and technically sound. Deploying AI in a lower-risk environment, where mistakes carry less severe consequences than in an emergency dispatch center, allows the technology to be calibrated, tested, and refined before expanding its scope. This kind of gradual approach is exactly what artificial intelligence experts recommend for public sector applications, where public trust is a critical asset that can’t be compromised by rushed implementations.
On the other hand, the fact that the city already took the bolder step of applying AI to 911 emergency calls shows that in some cases, operational necessity accelerates adoption before all ethical and regulatory questions are fully resolved. That’s not necessarily a mistake, but it is a signal that the public sector needs clearer frameworks for artificial intelligence governance in essential services.
As Dizengof put it, AI can serve communities in a better way and help save lives. But it needs to come with proper controls, where at the end of the day a human makes the final call on when and how AI will act.
New Orleans is, in a way, building this path while walking it. Transparency about the results, the mistakes, and the lessons learned from this process will be the deciding factor in whether the city becomes a model to follow or a case study on the risks of moving too fast. 🤖
Artificial intelligence is no longer a future promise for public services. In New Orleans, it’s already here, it’s working, and it’s redefining the relationship between the city and its residents — one call at a time.
