AI ‘consultant’ tool helps agencies speed up operational improvements
takasuu via Getty Images
The AI product underscores how tech can amplify human’s work not replace it, one official says.
Modernization and innovation are priorities for governments nationwide. But a major hurdle to getting there is identifying the challenges and obstacles in current practices that need to be fine-tuned.
For government agencies, process discovery and mapping can be brutal, as various staff and stakeholders come together to hash out how current systems operate with the end goal of identifying pain points that could be improved, said Brian Elms, co-founder of the company rvrwrk and former civil servant for Pennsylvania and the city and county of Denver.
But determining how agencies get their work done, who they work with, what tools they use and other details “can devolve into chaos [as] there can be a lot ego or some people lie because their boss is in the room,” he said.
Typical process discovery could take weeks just to come to a conclusion about how current systems work before agencies can even begin to implement changes to improve efficiency.
That’s where Coco, an artificial intelligence-enabled phone agent, comes into play. The tool was developed by the company rvrwrk and was trained on videos of a process improvement expert. With Coco, government users conduct an automated phone call during which they are asked about their workflow and operations over the course of 10 to 15 minutes.
Rather than offer advice or tell its users what to do, Elms said, Coco is “letting you tell the story” so it can produce resources like process maps, standard operating procedures and checklists to help inform governments’ operational improvements.
For South Bend, Indiana, Coco helped the city’s Department of Innovation and Technology modernization initiative move faster. Officials wanted to replace a software system used by code enforcement authorities to resolve maintenance and functionality issues, said Denise Reidl, the department’s chief innovation officer.
“Because of that digital transformation attitude, we [needed] to audit all of our processes,” she said. That meant reviewing how the software worked and where it faced difficulties, such as batch printing citations, and how municipal staff, the IT department and vendors interacted to get that work done.
The process was slow and complicated, often pulling code inspectors from their work in the field to sit down for hours at a time to review their operations or requiring numerous conversations and double checks of details, she explained.
Using Coco, the agency was able to audit nine different code enforcement processes in less than four hours across five employees. Faster, more efficient process discovery and mapping has major implications for agencies in the long term, Reidl said. By getting through the bureaucratic steps quicker, agencies can, for instance, innovate their services and systems sooner.
“These types of changes are always going to be time consuming, and any pressure that can be alleviated … is so helpful,” Reidl said. The AI tool helped the agency embark on its modernization journey “from square 10 instead of square one,” she said.
Agency staff still have to edit Coco’s output, meet with stakeholders and develop solutions, which highlights the need for human validation when using AI tools, Reidl said.
“One of our biggest learnings from [using Coco] is [that] Coco doesn’t replace process innovators in government,” she said. “It’s about elevating and improving and making the human process more efficient.”