Bringing AI to the Public Sector

Imagine a city where a resident posts on social media about a hazardous road condition, such as a large pothole, and tags the local Department of Transportation. That city’s social media monitoring system would instantly detect the post, process the information, and automatically generate a case management ticket.  

That ticket, in turn, would trigger alerts to the appropriate authorities and generate automated responses to the original post to provide status updates.  

“Encountering this idea sparked the question: Why isn't this already a reality?” Vishal Shrivastava (MSAI '23) said. “We possess the necessary technology to make it happen. What's hindering its implementation?"  

Tyler Technologies helps the public sector use technology to manage everything from court records and street maintenance to tax appraisals and public school transportation. To devise a system that uses artificial intelligence (AI) to curate citizen complaints, graduate students from Northwestern's MBAi Program — a joint-degree program offered between Northwestern's Kellogg School of Management and the McCormick School of Engineering — and Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program worked together to create a potential solution.

Four students – two from MBAi and two from MSAI – worked with Tyler on a pilot project to develop a solution using Fairfax County, Virginia, as a case study that would use AI to highlight urgent complaints.  

“Historically, the public sector has shown a cautious approach to advancing technology, with AI being no exception,” Shrivastava said. “Projects like Tyler's serve as opportunities to establish AI's credibility and rebuild governments' trust in it, reframing it as a beneficial technology rather than a threat.”

Shrivastava partnered with Yashika Garg (MBAi '23), Harry McGraw (MBAi '23), and Aanchal Sahu (MSAI '23) on the capstone project. The quartet was excited by the scenario Tyler presented.

"Tyler Technologies wanted to leverage AI to enhance the relationship between local governments and their communities," McGraw said. "This concept jumped off the page as an AI use case that could be readily envisioned and could result in a positive social impact." 

One of the students' responsibilities was to figure out a plan to overcome implementation obstacles.

One challenge was realizing residents don't follow a single route for complaints, instead turning to different social media platforms rather than official channels to report problems. 

That meant the team’s solution needed to focus on culling insights from multiple platforms — such as Facebook, X (formerly known as Twitter), Reddit, and Nextdoor — using natural language processing models.  

Another obstacle was identifying what posts were legitimate problems and what were more general complaints.   

“The idea was to build a solution using the AI techniques we learned in the program to distinguish between unusable sarcastic or hate speech and genuine complaints with sufficient information to take meaningful actions,” Shrivastava said. “The proposed solutions aimed to streamline the complaint submission process, improve transparency and accountability, enhance citizen satisfaction, and ultimately strengthen trust in local governance.”  

The students capped their capstone experience by presenting their recommendation directly to Fairfax County officials, a bonus the students appreciated.

"So often in school projects you work on something and never get to see the end user's true reaction, but to be able to present these ideas to local officials was a wonderful experience," McGraw said. "You simply cannot understand their pain points until speaking with them, and by having that connection throughout the project lifecycle, we were able to give them something we knew they wanted, rather than something we hoped they wanted."