Main an AI deployment in a corporation isn’t any small feat for any CIO, but it surely’s an entire totally different animal when that group is a metropolis.
Living proof: Mark Wittenburg, CIO of the Metropolis of Raleigh. His IT staff is spearheading an bold effort to use AI in areas starting from managing city progress and addressing local weather change to fixing potholes sooner. Wittenburg, talking at Gartner’s latest IT Symposium/Xpo occasion in Orlando, Fla., mentioned the aim is to not solely make metropolis providers smarter but additionally additional set up Raleigh as a part of a “good area.”
“Now we have to be modern,” he mentioned, noting that expectations are excessive. The realm’s constituents are on the lookout for the identical seamless expertise experiences they get from personal corporations.
Certainly, innovation is a part of the area’s DNA. Raleigh, the capital of North Carolina, anchors the brainy Analysis Triangle Park — the biggest analysis park within the U.S. and residential to company places for Cisco, IBM, Dell, Lenovo, and different tech giants. Wittenburg mentioned the town’s AI initiative is about enhancing providers for its 500,000 residents, whereas drawing on the experience of the area’s universities and schools.
Partnerships with the native universities in Raleigh — North Carolina State College, Shaw College and Meredith School — embrace working carefully with college college students, Wittenburg mentioned, sparking “some superb concepts and easy methods to remedy some real-world challenges that we’ve got.”
Breaking Down Silos to Streamline Workflows
The drive for IT innovation did not begin with AI. Six years in the past, earlier than trendy massive language fashions and AI applied sciences emerged, the Raleigh IT staff was already making an effort to work in a extra modern approach. The IT staff got down to look at siloed purposes and processes throughout varied metropolis departments to see the place these processes could possibly be streamlined, Wittenburg mentioned.
“How will we take these siloed workflows, like hiring any individual, after which what are the items these workflows contact alongside the best way?” he mentioned, for instance. “Then we begin to streamline these by sharing knowledge and sharing workflows.”
Testing AI Internally Earlier than Public Rollout
The emergence of generative AI triggered one other alternative to enhance the best way issues labored. The IT groups’ AI specialists began utilizing generative AI (GenAI) internally — Wittenburg did not need to instantly launch it to the general public ought to a significant situation come up. “There’s a few good examples of the place that is gone unhealthy,” he mentioned.
By beginning the AI journey internally, Raleigh’s IT staff was capable of start enhancing its personal processes first and create a extra environment friendly troubleshooting workflow. Raleigh is utilizing ServiceNow’s AI platform; Wittenburg mentioned that in consequence, 95% of assist tickets at the moment are auto-summarized and accredited, and tickets are cleared 66% sooner. By addressing tickets sooner, the IT staff can be capable of filter out service backlogs faster.
Enhancing Metropolis Providers with AI
When contemplating easy methods to use AI with public-facing processes, Wittenburg took a “crawl, stroll, run” strategy from ideation to manufacturing. That begins with asking a number of questions on which varieties of AI applied sciences take advantage of sense for the town of Raleigh, what enterprise worth AI can ship, and what the dangers are ought to one thing go fallacious, Wittenburg mentioned.
The primary software of a public-facing AI concerned Raleigh’s water system.
“We digitized all of the data, introduced all that in, after which constructed a mannequin utilizing knowledge analytics, AI, and generative AI to have the ability to begin asking questions round our water infrastructure,” Wittenburg mentioned, noting that the town’s paper data dated again to the 1700s.
The AI mannequin might then be developed and used to establish and predict potential failures within the water infrastructure. As well as, the strategy allowed for the creation of a closed mannequin utilizing Raleigh’s personal water knowledge with an internally constructed AI engine, which is a low-risk methodology for testing AI earlier than deploying it extensively, Wittenburg defined. Metropolis staff might additionally confirm the AI’s predictions and enhance the mannequin’s forecasts over time.
Simplifying Authorities Complexity
Subsequent, the IT staff built-in ServiceNow’s AI platform with the town of Raleigh’s public chatbot to implement “agentic items to it,” Wittenburg mentioned, referring to AI expertise able to autonomous resolution making.
“It was superb, as a result of now it began to make use of the information articles, but it surely additionally began to dig into these closed tickets to different workflows, and to begin to put all that collectively. It is simply superb watching it be taught issues that we did not know we had been doing,” he mentioned.
Whereas studying internally has been crucial, the final word aim of deploying AI is to deliver worth to the neighborhood, Wittenburg mentioned. In that vein, Raleigh’s IT staff plans to first use agentic AI to enhance conventional metropolis providers, corresponding to simplifying the method of figuring out and fixing potholes.
Fixing a pothole appears simple sufficient, but it surely entails a lot of steps, Wittenberg mentioned, beginning with figuring out if it is on a metropolis, county or state avenue — three totally different companies with separate telephone numbers. Individuals simply need the pothole mounted, the issue solved.
“What if we might get all these AIs and people brokers and all these providers to cover the complexity of presidency from you all,” he mentioned.
Imaginative and prescient Zero: Agentic AI for Safer Transportation
One other use case the Raleigh IT staff is engaged on with AI is “Imaginative and prescient Zero,” utilizing machine studying at site visitors intersections to create fashions round car and pedestrian interactions. The aim is to succeed in zero transportation-related fatalities within the metropolis. For Imaginative and prescient Zero, Wittenburg’s staff partnered with Nvidia and Esri, a GIS mapping software program, to construct an AI mannequin that might decide if drivers had been going the fallacious approach or if a automobile narrowly missed a pedestrian, for instance.
The IT staff is starting so as to add GenAI capabilities to ask the AI mannequin to counsel enhancements at intersections based mostly on “close to misses.” The AI mannequin has made ideas corresponding to including a proper flip mild or implementing mild change delays, Wittenburg mentioned.
Steering for CIOs
In the end, when deploying AI for the neighborhood, Wittenburg mentioned it is necessary for constituents to have a constant expertise and scale back complexity when they should join with an area company.
His recommendation to different CIOs and IT professionals beginning out on their IT journey is to establish the governance and “guard rails” for utilizing AI, have a stable knowledge basis and “do not let concern maintain you again.”
