Wicked Decluttering

In a recent Global AI column From Red Tape to Green Tape: Decluttering the State with AI and Collective Intelligence, I described an experiment we ran at the Open Government Partnership Summit in Spain. Hosted by Political Watch and mySociety, a group of global public entrepreneurs were asked a deceptively simple question: how could the City of Boston improve its permitting process without weakening the protections those rules exist to provide?

That experiment in Spain surfaced a recurring insight: regulatory systems fail not because rules exist, but because they are organized around institutional expertise rather than human purpose. 

The challenge is not deregulation, but re-organization—and in Boston we will call that wicked decluttering: simplifying the user experience without weakening the underlying protections. Identifying what that experience should be is why we need to use engagement to complement machine learning.

This week, at the newly opened AMA restaurant—sister establishment to the James Beard Award–winning Comfort Kitchen—Michelle Wu and Santiago Garces launched the City of Boston’s redesigned permitting experience at boston.gov/permitting.

“We are here today because we’ve been sprinting to simplify complicated systems and make them easier, more convenient, and more affordable for residents and businesses,” Mayor Wu said. “Permitting is meant to get easier.”

Garces framed what is at stake. When someone opens a restaurant, installs solar panels, or remodels a home, “they’re putting their money behind a dream.” Yet for years, Wu added, pursuing that dream required not just a business plan, but “becoming an expert in city bureaucracy just to figure out where to start.”

That mismatch between how governments organize rules and how people understand their goals was visible in the data. In a citywide survey, 47 percent of respondents said the most frustrating part of permitting was not knowing what information to trust or where to begin. As Garces put it, “We organize permitting around regulatory expertise—short forms, long forms, departments. But that’s not how people think. People are trying to redo a bathroom. They’re trying to open a restaurant.”

What are you trying to do and how can we help?

To realign the system around those intentions, the City partnered with AI for Impact Fellows at the Burnes Center for Social Change. Using AI to analyze 25 years of historical permit data, the team identified 260 distinct permitting experiences, clusters that reflect real-world intent rather than administrative categories.

“That analysis helped us answer a deceptively simple question,” Garces explained. “Where do we start?”

The City began with the twenty most common requests such as replacing a window, remodeling a kitchen or bathroom and rewrote permitting guidance in plain language. Instead of asking residents to navigate agencies, the new system asks a different question: What are you trying to do, and how can we help?

Crucially, those guides were not generated automatically, but co-produced by city staff across departments, using resident data and frontline expertise to iteratively improve clarity and accuracy.

Behind the scenes, AI is playing a second, quieter role. Tools developed in partnership with AI for Impact are helping staff in Inspectional Services, the Licensing Board, and related departments translate dense regulatory requirements into clear, consistent web content. A third, experimental use applies AI to improve search, so residents can find the information most relevant to their situation more quickly.

Pointing to AMA’s newly licensed bar, Garces described the effort as “translating deep regulatory expertise into guidance people can actually use.”

The city should be supporting them, not standing in the way

The result is not deregulation, but re-design. As Mayor Wu put it, “When people enter our permitting process, the city should be supporting them—not standing in the way.”

Boston’s permitting overhaul illustrates what responsible AI in government can look like: not automation for its own sake, but analysis and translation in service of human judgment, public value, and trust. Decluttering the state, it turns out, is not about removing rules, it’s about reorganizing them around the lives people are actually trying to live.

Wu situated the permitting work within a broader vision for neighborhood life: “Our goal is to make Boston the best place to do business, the easiest place for families and entrepreneurs to have what they need and to thrive,” said Mayor Wu. “I also know the incredible community that can only be built when there are special places like this—places where people come together, celebrate life’s biggest moments, reconnect with friends, or even just sit down for a hard conversation. Our restaurants and small businesses provide that across every neighborhood in the city.”

From left to right: Julia Guttierez, David Fields, Beth Simone Noveck, Shin-pei Tsay, Aesclinn Donohue, Santiago Garces