Introduction
In the summer of 2024, I was selected as the Bloomberg Harvard Summer Fellowship for the City of Newport News, Virginia.
The project was simple: how might generative artificial intelligence transform city services?
Opportunity identification
Drawing on my service design background question, I devised a strategy of pairing user needs with AI capabilities. "Ask not what you can do with AI, but what AI can do for you."
To understand user needs, I planned and executed 41 interviews and site observations with 88 stakeholders across most city departments.
To understand AI capabilities, I conducted desk research on consultancy reports/frameworks and had conversations with experts in the field.
Among my most memorable sessions included a 1-hour focus group with the Fire Chief, Fire Marshall, five Assistant Fire Chiefs, and three support staff all in the same room! I also got a tantalizing view of the city's 911 call centre, interviewed the mayor, and learned bits and pieces of so many other city operations, from libraries to vehicle maintenance to the graphic design team.
Use-case prioritization
In two months' time, I identified 95 use-cases for GenAI, then prioritized them across 5 desirability criteria and 5 feasibility criteria through follow-up consultations with stakeholders and executives. This detailed, rigorous, and co-creative approach inspired confidence in my recommendations.
Implementation roadmap
To make the long list of use-cases more approachable and actionable, I defined a roadmap consisting of:
—13 enterprise implementation pods (to be spearheaded by the enterprise IT department in 5 sequential phases); and
—10 vertical implementation pods (for individual departments to explore in parallel at 3 levels of priority).
Final deliverables
At the end of the summer, I submitted a 22-slide final report and 165 slides of appendices. Contents included:
—process documentation, including decision-making rationale, prioritization criteria, lists of stakeholder engagements, and desk research insights;
—23 departmental profiles to summarize stakeholder perspectives on AI, both optimistic and cautious;
—23 implementation pod placemats connecting themed groups of use-cases into buckets sharing similar implementation processes and skillsets; and
—95 use-case one-pagers containing performance indicators, prioritization criteria, and implementation considerations.
Some excerpts provided below...
It's hard to explain, but to me as a photographer, the Virginia sun shone in a captivatingly luminescent, strikingly golden glow I've never seen anywhere else....
