Data-Smart City Pod – Episode Summary
Episode: How GenAI Can Actually Boost Public Sector Creativity
Date: October 29, 2025
Host: Stephen Goldsmith, Bloomberg Center for Cities at Harvard University
Guest: Mike Sarasti, Entrepreneur, Former Chief Innovation and Information Officer, City of Miami

Episode Overview

In this episode, Stephen Goldsmith engages in a forward-thinking conversation with Mike Sarasti about how generative AI (GenAI) can unlock new realms of creativity in public sector processes. Drawing from Sarasti’s deep experience with government innovation and technology, they explore how GenAI can go beyond mere efficiency to foster breakthrough thinking, enhance process mapping, democratize AI’s benefits for city managers, and more effectively engage both public servants and communities in city transformation.

Key Discussion Points & Insights

The Role of Creativity in Government & AI

• Both agree that public sector “efficiency” is often about tweaking old systems rather than reimagining them, and that creativity is essential for real breakthroughs.
  • Sarasti emphasizes the importance of protecting space for creativity and human “play,” even as automation and AI increase efficiency.

“Problem solving is an act of creation … Preserving that weight on craft and the spark that comes from creation … is really important to preserve.”
— Mike Sarasti (01:22)

• Using a musical analogy, Sarasti calls for government work to retain a sense of “jamming,” collaboration, and exploration, not just “shaving off seconds” with new tools.

“The breakthrough stuff does come out when you have sort of some freedom to do it … create more space for play, more space for actual human collaboration. Because you're offloading all of these tedious things …”
— Mike Sarasti (03:54)

Mapping Processes at Machine Speed: Riverwork’s Approach

• Sarasti recounts his transition from open data advocacy to “machine readable process,” shifting from collecting datasets to capturing and mapping the actual steps and interactions that drive city operations.
  • He explains how his startup, Riverwork, uses GenAI to rapidly collect process information through conversational interviews and automate what once took months or years.

“We were able just recently to map out 100 processes in a little over 30 days. If you would have asked me to map out 100 processes as a CIO, I would have told you that we've got a three year project on our hands …”
— Mike Sarasti (10:21)

• Riverwork’s method leverages AI-driven agent interviews, enabling many simultaneous contributors and capturing multiple perspectives for a composite process map.
  • This decentralizes knowledge collection, captures “the stories,” and avoids the pitfalls of the loudest voices or groupthink.
  • Converts conversational data into structured, machine-readable formats (e.g., JSON), making organizational knowledge flexible and reusable. (12:57)

Tackling Interagency Complexity

• Goldsmith presses on how to map and improve interconnected processes that span multiple agencies—critical for true customer-centric government (e.g., multi-department permitting).

“How do you process map 10 agencies or six agencies into a combined customer facing solution?”
— Stephen Goldsmith (12:34)

• Sarasti explains that their AI interviews naturally surface cross-agency dependencies as staff describe real-world workflows. Once transcripts are gathered, GenAI can analyze relationships—such as which staff or systems are mentioned frequently—building a multi-perspective network of process nodes.

“You can establish those relationships. … Now that you've got these transcripts, turns out you can ask a lot of questions that I hadn't even thought about asking at the beginning … you can establish the nodes … and make some of those connections.”
— Mike Sarasti (13:32)

Democratizing AI for Mid-Managers and Data Literacy

• Goldsmith asks how to empower not just data teams but mid-level managers to use GenAI to investigate recurring problems, like repeated pothole repairs.

“How do we partially democratize the use of generative AI to access data? … We want to have more curiosity by mid managers about why they're filling the same pothole 10 times. How do we take the next step?”
— Stephen Goldsmith (14:35)

• Sarasti advocates for promoting “prompt engineering” as an essential new skill, enabling managers to guide AI to find patterns and generate insights—even without deep technical expertise.

“You can take a document that's a PDF and you can have a conversation with it and beef up your ability to prompt it. I think those skills are really, really important and very valuable for a performance management.”
— Mike Sarasti (15:31)

• He warns against using GenAI solely for “automated” answers via shallow prompts; instead, encourages thoughtful, multi-step inquiry for richer, more actionable results.

Leadership, AI Vendors, and the Importance of Human Knowledge

• Sarasti cautions leaders to be skeptical of vendors promising magical solutions if you “just give us your data,” since most vital knowledge remains in people’s heads and legacy systems weren’t designed for easy extraction.

“A lot of these data systems … it’s a sliver of what’s happening … you’re still coming back to humans. The knowledge is still trapped in people’s brains.”
— Mike Sarasti (18:17)

• Leadership should:
  • Focus on amplifying—not replacing—human creativity.
  • Use GenAI to enrich traditional data sets with narrative, context, and nuance.
  • Be wary of “automagic” pitches and insist on approaches that blend technology with the lived experience of employees and communities.

Community Engagement & Next Steps

• On whether community members are involved in redesigning processes, Sarasti admits much more can be done, but foresees conversational AI playing a vital role in gathering and analyzing narrative feedback from constituents.

“You can put people in a stance to tell those stories fully and then use some of this tech to look for patterns …”
— Mike Sarasti (20:23)

Accelerating AI-Powered Transformation in Cities

• To speed up positive change, Sarasti says cities should provide broad access to GenAI tools, letting staff learn their quirks and possibilities firsthand.

“We need to start putting these tools in people’s hands and encouraging people to not let go of their agency … You want people to think about it again with curiosity and a process of discovery.”
— Mike Sarasti (22:06)

• Experience is key: only by using these tools can managers and leaders develop the necessary intuition about their capabilities and limitations.

Memorable Quotes & Moments

• On government creativity:
  “If we suck out the creativity of government work … fundamentally, we’re all creative beings … problem solving is an act of creation.”
  — Mike Sarasti (01:22)

• On making space for breakthrough thinking:
  “The breakthrough stuff does come out when you have sort of some freedom to do it … and you're, you're sort of set up with these great tools.”
  — Mike Sarasti (03:54)

• On the limitations of government data systems:
  “The data you have … it’s only … we’ve been working … with like, what we got. It's usually like the data that's easier to get to.”
  — Mike Sarasti (18:17)

• On the risk of overpromising AI magic:
  “Be cautious of the people that are promising you all this stuff that AI can sort of do and lean into the things that are really … more opportunity for human creative stuff.”
  — Mike Sarasti (19:34)

Timestamps for Noteworthy Segments

| Timestamp | Topic | |------------|-------------------------------------------------------------------| | 01:22 | Creativity’s role in AI and government process transformation | | 03:54 | “Play” vs. optimization in reengineering government workflow | | 05:40 | Sarasti’s career and journey in Miami innovation & technology | | 08:04 | Machine-readable process and rapid process mapping with AI | | 10:21 | Case study: Mapping 100 finance processes in 30 days | | 12:57 | Mapping and analyzing cross-agency, customer-centric flows | | 14:35 | Democratizing AI for mid-manager curiosity and data literacy | | 15:31 | The core skill: prompt engineering for GenAI | | 18:17 | What leaders should and shouldn’t expect from AI in government | | 20:23 | Involving communities in redesign: gaps and future potentials | | 22:06 | Advice for accelerating AI-powered transformation in cities | | 23:06 | Closing banter about AI as a “confidence booster” |

Conclusion

This episode offers a nuanced, actionable roadmap for city leaders and innovators seeking to harness GenAI not just for incremental gains, but for ambitious, human-centered transformation. Sarasti’s insights highlight the need to preserve creativity, democratize powerful tools through prompt literacy, use AI to unlock hidden organizational knowledge, and above all, keep the human experience at the core of technological progress in the public sector.