Ville de Laval speeds up community response systems with an Azure AI solution

One of the largest cities in Quebec, Ville de Laval, is focused on developing smart city initiatives. One such project is the infusion of AI and machine learning into its 311 non-emergency hotline. The system, which fields nearly 200,000 inquiries a year, has begun using a virtual agent to both speed citizen–agent interactions and answer the more basic inquiries on its own. Built with capabilities in Microsoft Azure AI and Microsoft Power Platform, the system was also recently used to conduct citizen consultations, and it may soon be expanded to the city’s 911 emergency response system. Ville de Laval expects that it will pay for itself within its first year of operation.

“By automating a large number of calls, and through the optimization of our operating costs, we expect that our Azure AI solution will have paid for itself by year’s end.”

Patrick Taillon, IT Analyst and Innovation Strategist, Ville de Laval

Download video transcript

Located just northwest of Montreal, Ville de Laval is the third-largest city in Quebec and the 13th largest in Canada, with a population of more than 400,000 people. A modern, growing city, Laval puts a strong emphasis on providing its citizens with a high quality of life through cutting-edge infrastructural initiatives. These initiatives are each part of a process of digital transformation architected by Marc Campeau, Innovation and Technologies Director at Ville de Laval. The introduction of AI to the city’s 311 system is a prime example of the work being done.

As in many other cities around the globe, the 311 system in Laval is set up to field questions and complaints about non-emergency issues that citizens might be experiencing. The sudden appearance of graffiti or a missed trash pickup are both commonplace 311 issues.

For Christine Lacroix, manager of the 311 call center, the introduction of the virtual service has had a positive impact on the citizen experience, not only in sending requests through, but also in building their confidence that the requests will be well understood and executed in a timely manner. “By removing the clerical task for the agent to enter the request in the system, it naturally creates space for the citizen to build a stronger relationship with the agent, and, by extension, the city.”

For a city the size of Laval, the fielding of these calls can become a sizeable undertaking. On average, Laval handles 645 calls a day. If you also account for in-person requests and emails, the city’s 311 operators respond to more than 250,000 requests annually. With so much going on, the manual recording of each call can sometimes be left incomplete. The push to augment the system with AI began in 2019. “We wanted to add value for each person contacting us through the 311 system,” explains Patrick Taillon, IT Analyst and Innovation Strategist at Ville de Laval.

Laval was already using the Microsoft 365 productivity cloud for multiple workloads, making the adoption of Dynamics 365 and Azure a simpler process than adopting a third-party solution might have been. Nonetheless, the city conducted a multivendor survey that even included the creation of its own AI assets through open-source technologies. “Through the vetting process, I quickly realized how simple it would be to create a very high level of interoperability between Azure and our Microsoft 365 environment,” says Taillon. “And with so much of our infrastructure already in the Microsoft cloud, developing a solution that uses capabilities in Azure AI is made much simpler, as it would reside in the same tenant.”

Thrilled with the results of the 311 service integration, Ville de Laval is already expanding its use of the Microsoft solution to other city services. For example, Ville de Laval recently used Microsoft Azure Cognitive Services in a December 2020 virtual citizen consultation session with positive results.

As an integral part of the Ville de Laval’s Public Consultation and Citizen Participation Policy adopted in April 2021, the public consultation team in the Communications and Marketing Department is responsible for gathering collective intelligence from Laval’s citizens and inform the decision process. According to Gabrielle Immarigeon, who leads the Public Consultation Division, one of the main concerns is "that participants are able to access and understand the information about the projects, and that it is accessible to as many people as possible, including those with functional limitations and language differences. This is an important aspect in our dealings with citizens because when we ask for their opinion, we want to make sure that their comments are informed and based on a good understanding of the projects."

As such, the virtual agent transcription service provided by Azure Cognitive Services is useful in a virtual consultation meeting in that it displays real-time transcription in French in Microsoft Teams. This ensures that people with hearing disabilities as well as those who wish to listen to the presentation while simultaneously reading the text gain a better understanding of the proposed projects, as well as contribute actively to the discussion.

Driving collaboration between humans and AI

While Laval worked with the technology experts at Gestisoft, a member of the Microsoft Partner Network, to deploy Dynamics 365—and use its customer relationship management functionalities to create an improved database of the issues citizens have consulted the 311 system about—Taillon began implementing the city’s new AI solution. The solution incorporates Azure Cognitive Services, Azure Machine Learning, Microsoft 365, and Microsoft Power Platform. The goals of this deployment were to automate the delivery of call summaries, classifications, and operational analytics to both Dynamics 365 and Power BI. “By generating reports in Power BI, we can begin to understand which calls our agents are fielding most frequently,” says Taillon. “Automating the way we provide that information, we can substantially improve the quality of the service we provide.”

The introduction of AI and machine learning to its 311 service helps Laval aggregate calls more efficiently. Quick, information-seeking calls can be managed by the city’s virtual agent, reducing wait times and allowing employees to respond to complex requests sooner. Already implemented, the solution is expected to automate 20,000 calls by the end of its first year of operation. “By automating a large number of calls, and through the optimization of our operating costs, we expect that our Azure AI solution will have paid for itself by year’s end,” observes Taillon.

“Contrary to what most could imagine, the AI is meant to aid our 311 responders, not replace them,” Taillon points out. “As we rolled out the solution, we made a point to sit down and ask employees how it might help them better accomplish their most important work—those tasks that we will always need a person to take charge of. Nowadays, I believe they take pride in being some of the first people working for Ville de Laval to utilize AI in this way.”

In addition to the possibilities of responding to many 311 calls with automated replies, the city is using Azure Cognitive Services with machine learning to recommend citizen request type to the agent during calls between citizens and agents. The system then surfaces relevant information for the agent, reducing the time spent looking up records and service schedules. The solution also provides real-time transcription across multiple languages, automatically generates call summaries for agents, and provides sentiment analysis for future reference. For managers, the system can also provide helpful insights into areas where agents can benefit from additional training, or if they are fielding the sorts of calls that might lead to excess stress during their workday.

Delivering faster results when they matter most

Laval is considering expanding his Azure AI–driven solution to more areas, including the emergency response 911 system, because the solution is ready for it. “We want our 911 system to be capable of detecting numeric information and surfacing that to our agents,” says Taillon. “If people send a text or a video to 911, for instance, we are working on a system that will be able to understand that information and rapidly contact an agent while providing as much context as possible.” Laval foresees its virtual agents as a key part of its response network, especially in the event of a large-scale emergency.

“We have seen events—in the United States, for instance—where a violent event occurs and the 911 system is quickly overwhelmed by thousands of calls in the space of a single minute,” explains Taillon. “During events like this, agents have actually had to turn to social media to develop an understanding of what is happening in their city. With a virtual agent aiding responders and delivering them actionable data sourced from hundreds of concurrently filmed videos, we are hoping that we will be able to guide people away from danger faster than ever, while assuring our first responders have the information they need to arrive on site prepared.”

The work completed around the city’s 311 system has already garnered high praise from the Quebec Municipal Informatics Network, which gave the solution the 2020 Excellence Award—the top honor the group can bestow. Other municipalities around Quebec could also be interested in these systems. “It is not as if the benefits we’re looking to create are limited to Laval,” says Taillon. “Population centers across Quebec, many of whom already use Microsoft 365, could implement these same solutions very rapidly by pairing Azure Cognitive Services with their existing Power Automate clients, and greatly benefit from it in terms of performance, and employee and customer experience.”

Find out more about Ville de Laval on Twitter, Facebook, YouTube, and LinkedIn.

Photo credit in order of appearance: Sylvain Majeau, Annie Poirier, Annie Poirier et Jany Tremblay.

“By generating reports in Power BI, we can begin to understand which calls our agents are fielding most frequently. Automating the way we provide that information, we can substantially improve the quality of the service we provide.”

Patrick Taillon, IT Analyst and Innovation Strategist, Ville de Laval