Detailed Summary: “From Reactive to Preventive: How AI Transforms Public Works”

Podcast: Data-Smart City Pod
Episode Date: April 1, 2026
Host: Stephen Goldsmith, Professor of Urban Policy at the Bloomberg Center for Cities, Harvard University
Guests:

• Daniel Pelaez, CEO & Co-founder of Civil
• Khalil Luizi, Executive Director, Public Innovation Institute
• Mike Dennehy, Former Commissioner, City of Boston Public Works

Episode Overview

This episode explores how artificial intelligence (AI), particularly visual analytics and automation, is transforming the way cities manage and maintain public infrastructure. The conversation centers on moving from a reactive, complaint-driven approach (e.g., 311 calls) to a proactive, preventive model powered by continuous data collection, analysis, and actionable insights.

Guest Introductions & Backgrounds

• Daniel Pelaez (00:56): Co-founder of Civil, a company using robotics and AI to help cities manage physical infrastructure—drawing inspiration from his early experience in public works and studies in autonomous vehicle technology.
• Mike Dennehy (01:55): Former Boston Public Works Commissioner, with decades of experience in city operations and public works management, now supporting Civil’s team.
• Khalil Luizi (02:20): Economist focusing on the intersection of technology, innovation, public health, and urban infrastructure, and Executive Director of the Public Innovation Institute.

Key Discussion Points & Insights

1. Citizen-Reported vs. Algorithmically Detected Issues (03:10 – 05:44)

Khalil Luizi describes his recent research comparing traditional 311 citizen reports to AI-enabled detection in Boston:

• Highly engaged neighborhoods file more 311 reports, typically resulting in more governmental resources allocated their way.
• Lower-income and immigrant communities are underrepresented in 311 data and thus risk being underserved.
• AI systems like Civil’s can detect issues—like hairline pavement cracks—that residents may not notice or report, increasing both the scope and equity of infrastructure maintenance.

Notable Quote:
“Engagement was lower in some of those [marginalized] areas. The technology was picking up things that people ordinarily wouldn't resort to.”
—Khalil Luizi (05:26)

2. How AI & Visual Analytics Enable Proactive Maintenance (05:44 – 07:53)

Daniel Pelaez explains the traditional manual, labor-intensive inspection methods (clipboards, ruler, paper records), highlighting inefficiency and limitations. He contrasts this with automated, sensor-based mapping and AI’s consistent, repeatable data:

• Sensors (cameras, 3D mappers) enable objective identification of minor infrastructure issues before escalation.
• AI removes subjectivity, providing cities with earlier warnings about potential problems, leading to considerable savings and better planning.

Notable Quote:
“These sensors...are able to produce very consistent, repeatable results. So, so they can identify that hairline fracture and how much it's maybe widening over time so they can proactively go fix that thing before it becomes an issue.”
—Daniel Pelaez (07:32)

3. Community Engagement, Data Democratization & Advanced Analytics (07:53 – 10:16)

• Large Language Models (LLMs) vs. Visual Language Models (VLMs): Daniel deep-dives into technical advances, emphasizing how complex visual data can now be indexed and queried similarly to text, unlocking a searchable “infrastructure index” for cities (08:15 – 09:38).
• Augmenting with Resident Contributions: Residents’ photos/videos could enhance the indexed data, facilitating endless use cases via natural language queries.
• Mike Dennehy (10:16 – 12:25) reflects on Boston’s historical data practices and how real AI-powered insight would have turned public works from reactive to proactive—enabling better budget justification and smarter long-term infrastructure planning.

Notable Quotes:

"Someone with natural language could say, 'Show me the dangerous intersections based on our prior data...with maybe faded pavement markings and high traffic.' ... It's possible."
—Daniel Pelaez (09:26)

“This technology is a game changer. It would allow these departments to analyze and process data so quickly and precisely that [they] go from reactive to proactive very, very quickly.”
—Mike Dennehy (11:38)

4. Democratizing Spatial Data and Departmental Transformation (12:25 – 14:48)

• Advanced platforms let managers ask questions like “Show me repetitive potholes, drainage issues, or tree issues” via natural language—this functionality is already available, awaiting broad adoption by middle and upper management.
• The power for cross-departmental use is emphasized; any agency managing public assets can benefit.
• Daniel cites Boston’s CIO, Santi Garcez, as a pioneer, describing internal agents that let users “talk to infrastructure” for instant, actionable reports.

Notable Quote:
“We built an MCP [Management Command Platform] for Boston's infrastructure. Santi and his team can now ... just natural language say, ‘Hey, there’s some flooding ... help me find all the manhole covers and the catch basins ... overlay that with resident complaints. Give me a report I can give to the mayor.’”
—Daniel Pelaez (14:23)

5. Ensuring Equity in AI Systems for Public Works (14:48 – 16:34)

Khalil pushes for “multimodal” approaches—using both AI and human reports—to avoid bias and ensure ground truth:

• Need rigorous benchmarking so the ground truth is correctly established and not based solely on potentially biased legacy data.
• Combining citizen input, technology, and external data maximizes both accuracy and equity.

Notable Quote:
“We have to ensure that we're benchmarking the systems against established ground truth...It has to be a combination or multimodal approach to designing some of these systems.”
—Khalil Luizi (15:17)

6. Procurement, ROI, and Future Adoption (16:34 – 17:03)

Daniel Pelaez addresses challenges cities face in acquiring such technologies:

• Uncertainty around purchasing models—should cities buy solutions outright or subscribe as a service?
• Cost-benefit: AI-powered systems are markedly cheaper and more effective than manual inspection.
• Partnership ecosystems and cross-agency collaboration facilitate smoother rollout.

Outlook:
“Infrastructure intelligence ... is going to start to become commonplace within governments. ... 12 months from now, I think we’ll have unlocked some things for agencies we wouldn’t have thought of before.”
—Daniel Pelaez (17:36)

Memorable Moments & Quotes

• Mike Dennehy on Future Departments (18:47):
  “I see middle managers walking into budget hearings with folders that include intelligence reports ... that specifically call out things that may not have come into 311. … This technology has addressed and can provide preventive maintenance or more long-lasting effects from a repair basis.”

• Khalil Luizi on Broader Applications (19:44):
  “This data is very rich for informing decisions across sectors ... infrastructure data can be transformed into epidemiological signals and assessing environmental risk factors.”

Timestamps for Key Segments

• (03:10 – 05:44): Comparing citizen vs. AI reporting—implications for equity
• (05:44 – 07:53): How AI and sensors revolutionize infrastructure management
• (07:53 – 09:38): Indexing the physical world, mixing resident input and AI
• (10:16 – 12:25): Departmental impact—moving from reactive to proactive
• (12:25 – 13:40): Democratizing data—how close are we?
• (13:56 – 14:48): Real-world use: Boston’s internal infrastructure AI agents
• (14:48 – 16:34): Equity and ground truth in AI-based systems
• (16:34 – 17:03): Procurement and implementation challenges & models
• (18:47 – 19:44): Vision for city departments and cross-sector insights

Conclusion

The episode highlights a paradigm shift in public works: cities no longer need to wait for resident complaints or rely on sporadic manual inspections. AI-driven tools make preventive maintenance and equitable service possible, while also streamlining budgets, supporting cross-departmental decision-making, and, ultimately, improving quality of life for all residents. As infrastructure intelligence platforms become more widespread, their measurable impact on cities will only grow.

Selected Quotes Recap:

• “The technology is able to pick it up.” —Khalil Luizi (05:24)
• “It would allow these departments to ... go from reactive to proactive very, very quickly.” —Mike Dennehy (11:38)
• “You can ask this thing anything and it's going to tell you exactly what's going on and what to do.” —Daniel Pelaez (13:56)
• “It’s obligatory for these middle managers to then make this the priority.” —Mike Dennehy (13:21)
• “It has to be a combination or multimodal approach ... [to] improve equity.” —Khalil Luizi (15:31)