AI to Listen Better
Alex Webb / Magnum Photo

What hundreds of hours of conversations with New Yorkers reveal about everyday life — and how cities can earn trust.

A 54-year-old woman in Bedford-Stuyvesant told us what she loves about her neighborhood: She can step outside her door and find Thai, Caribbean, Mexican and Greek food, all within a few blocks of her home. But remaining in the city she loves is increasingly difficult: “So even with the escalating costs, I’ll try to stick it out as long as I can. I love New York.” Her attachment to the city is shared by many New Yorkers who are increasingly uncertain whether they can afford to stay, or feel safe raising their families here.

A 33-year-old mother in Manhattan told us: “I would like to stay 'cause I was born here, but if the crime rate doesn't go down, I'm gonna have to leave because my kid's not going to be a victim to the street.”

In a few words, these New Yorkers captured both what makes the city so special and why staying is becoming increasingly difficult.

City leaders regularly collect, analyze and report on thousands of data points about life in New York — from 911 calls and 311 complaints to crime statistics, school performance, subway reliability, housing affordability, cost of living, homelessness and street cleanliness. But there’s a difference between how a city might look on paper and how residents experience it.

Crime can fall while high-profile incidents and the attention they attract leave residents feeling less safe. Graduation rates can rise while students feel unprepared for what comes next. Streets can meet official cleanliness standards while residents still encounter rats outside their homes. Consider what life has become like for a 67-year-old man in Brooklyn, who told us: “I don't even like to go out of my apartment. I wait until the middle of the night to throw out my garbage or get my mail. That's how bad it is in this place and how much I want to get out of here.”

It's often the daily struggles, fears and uncertainty that leave people feeling less secure and less confident that things are moving in the right direction. But it's also the small signs of progress that can give them hope. A pothole finally filled, a neglected building on your morning walk finally repaired or a rent freeze that helps you sleep better at night — these are changes people can see and feel, and that give them hope things are getting better.

Understanding how people experience their city, what makes them feel more or less secure, and whether they believe life is getting better is essential to measuring quality of life and building the trust that helps a city thrive.

Yet until recently, capturing these experiences, the signs of progress or decline New Yorkers recognize — in their own voices, systematically and at scale — has been difficult.

What we built

For nearly three decades, I've spent much of my professional life listening to people, trying to understand what they believe, how they feel and the experiences that shape their views. My work has combined quantitative surveys — what most people think of as traditional polling — with deeper conversations through focus groups and town halls. 

Surveys tell us how many people share an opinion or concern and how their views differ across communities and demographic groups. But it's often in the open-ended conversations I have using qualitative research methods, when we’re able to create space for people to feel comfortable opening up, that I learn the most. Sometimes it's an unscripted follow-up question in a focus group. Other times, it’s something someone volunteers that I never would have thought to ask about that changes how I understand an issue, a point of view, or the struggles someone is facing.

Advances in AI now make it possible to create a digital twin of myself — an AI interviewer that speaks in my voice, is trained in my interviewing techniques, and conducts in-depth, one-on-one conversations, asking follow-up questions much as I would in a focus group. This combines the strengths of both approaches: the scale and reach of polling with the depth and human connection of an open-ended focus group-style discussion. Since May, in partnership with Vital City, we have been putting this “qual at scale” approach to work in New York, using CrowdVox AI, a platform I helped develop. 

We recruit New Yorkers from established market research panels, inviting them by email to take part in an interview. Participation is voluntary, and while the people we speak with are not randomly selected, we make a concerted effort to hear from New Yorkers who reflect the city's diversity across age, gender, race and ethnicity, education, and neighborhood.

The AI interviewer then engages New Yorkers in one-on-one phone conversations about their experiences living in the city, asking follow-up questions much as a skilled moderator would. The responses are then stored in a secure database, where they can be analyzed, visualized and searched by topic, frequency, sentiment, emotions and other measures. The goal is to give various stakeholders like city officials, neighborhood associations, or even residents themselves the ability to explore the results and query the database to uncover patterns, compare experiences across neighborhoods and demographic groups, and ask questions that surface new insights, ideas and potential solutions. Since May, we’ve conducted more than 3,282 interviews, lasting a combined 587 hours, across all five boroughs and 174 of 177 (98%) residential ZIP codes. We track many of the same questions every day, while also rotating new ones every few weeks. The average length of our interviews is just under 11 minutes, with the longest running a half hour. The interviews and broader dataset — which we adjust each month to ensure the respondents reflect the composition of the city as much as possible — allow us to track how New Yorkers feel about their city, borough, and neighborhood across key measures, including economic security, safety and predictability, future and opportunity, health and well-being, neighborhood life, trust and governance, youth and education, and mobility and public space.

So far, we’ve learned three important lessons from this work.

Lesson one: People like being heard, especially when they can speak freely  

When asked to rate their overall experience with the AI interview, New York City residents with whom we spoke gave an average score of 4.5 out of 5 – and 80% of participants gave permission to share their voice and views publicly. “I felt heard, I felt listened to, the AI didn’t interrupt and let me finish my sentence,” a 29-year-old woman told us.

Making people feel comfortable and heard may sound simple, but too often, we fail to get it right. And for some, the experience with the interview went beyond feeling comfortable. It offered a sense that someone was genuinely interested in what they had to say.

“I never had a person that actually cared enough to understand my circumstances,” a 25-year-old woman told us.

Consider how local governments typically hear from its residents: a 311 ticket, a multiple-choice survey, a two-minute opportunity to speak in a community board meeting – they all ask residents to fit their experiences into whatever box the institution can process. AI-led interviews give New Yorkers the time and space to explain what they’re experiencing, in their own words, and allows for follow-up questions that help us understand why. 

Lesson two: Numbers only tell part of a city’s story

Two-thirds, or 65% of New Yorkers in a Vital City poll we conducted late last year — completely apart from the listening project I’ve just described — said they had cut back on groceries, transit or other essentials because of cost. That statistic tells us how widespread the affordability crisis is. But when we hear directly from New Yorkers about the sacrifices they’re making, the pressures they’re under, and the emotional toll of trying to get by, the numbers come alive and can inspire new ideas and solutions. City leaders also gain a clearer understanding of what residents are experiencing and what they need, giving them a stronger case for action within city government. Asked about the economy and his standard of living, a 52-year-old city employee in Brooklyn shared something that stopped me in my tracks: “I give plasma twice a week. … I’m always trying to, you know, hustle, as they say, to survive in New York City.” I can offer poll numbers about affordability, and the City can share charts and graphs on a dozen economic variables — but when a city employee tells us he’s selling plasma twice a week to keep his head above water, that's not easy to forget. It's the kind of data that can change how we understand a problem and create the urgency to do something about it.

Lesson three: Listening systematically helps us see what’s coming

Like a public opinion survey, a set of focus groups, or a community meeting, a single round of AI-enabled interviews is a snapshot in time. However, by asking the same questions, the same way, every day we’re able to identify subtle changes in the issues people raise, how they talk about them, and emerging concerns that might otherwise go unnoticed.

One of the clearest examples comes from responses to our question, “What is one thing you have seen or heard in the news recently that stuck with you?” During late spring and early summer, 5% of New Yorkers raised immigration or ICE-related topics. But by October, that number tripled to 15% and has remained elevated at 13% so far in November. Earlier conversations focused on immigration policy, but by fall, New Yorkers were increasingly talking about ICE and specific enforcement actions. And residents on both sides of the issue were increasingly factoring those concerns into how they evaluated City Hall.

We saw a different pattern around the city's rent freeze. In May, just 2% of residents mentioned it without being asked. Following the Rent Guidelines Board's June decision, that number jumped to 16% in July. But by October, when the freeze officially took effect, it had fallen to just 3%, where it remains today.

Together, these examples show how daily listening can help us recognize not just when an issue becomes more urgent, but how people's concerns, expectations, and reactions change over time. Understanding those differences gives city leaders a better chance to respond in ways that make residents feel heard, understood, and taken seriously.

A field taking shape

Ours is far from the only effort in AI to bridge the gap between people’s desire to be heard and the capacity of even the most willing institutions to listen.

Google’s Jigsaw initiative is working with cities to use AI to engage more residents, understand their concerns, and identify areas of common ground. In Bowling Green, Kentucky, nearly 8,000 residents contributed ideas that helped shape a 25-year community plan. Cities including Chattanooga, Tennessee, and Kitchener, Ontario, are now experimenting with similar approaches to inform decisions about public transportation and long-term planning. In Tokyo, the metropolitan government is using AI to help shape its long-term vision for the city in the 2050s. Through an initiative called “broad listening,” local officials collected residents’ views through social media, online surveys, mail and in-person interviews, then used AI to identify common concerns and organize ideas — in turn, making thousands of perspectives easier for policymakers to understand. 

Building a city that listens

New York can now make this kind of listening a foundational part of how the city operates. Engaging residents consistently and reaching deeper into communities where people often feel isolated or unheard, city leaders can better distinguish the signal from the noise, understand what matters most to residents, and recognize emerging concerns before they become crises.

We hope that what we're building is more than a new research tool. It's a form of civic infrastructure that can help connect the daily experiences of New Yorkers to the decisions that shape their lives. Its real value will depend on whether those voices help shape decisions — and whether New Yorkers can see the difference.

When people feel heard, trust has a chance to flourish. When leaders understand what makes life better or harder for residents, they can make more informed and impactful decisions and show residents their voices matter. And when New Yorkers feel connected to their communities, confident in their future, and hopeful, that's when New York is at its best.


Great! You’ve successfully signed up.

Welcome back! You've successfully signed in.

You've successfully subscribed to Vital City.

Success! Check your email for magic link to sign-in.

Success! Your billing info has been updated.

Your billing was not updated.