This is like a smart, conversational tour guide for Washington, DC’s open data. Instead of downloading spreadsheets and decoding columns, any resident or city staffer can just ask questions in plain English—“Where are the most traffic crashes?” or “How many affordable housing units were built last year?”—and the AI finds, summarizes, and explains the relevant data.
Traditional open data portals are hard for non-experts to use: they require data literacy, time to filter and download datasets, and technical skills to analyze them. This GenAI layer turns complex, siloed datasets into natural-language answers, expanding who can actually benefit from open data (residents, journalists, small businesses, internal staff) and reducing city staff time spent on ad-hoc data requests.
Combination of curated municipal datasets, domain-specific prompt engineering, and integration with DC’s existing open data portal and governance processes; over time, interaction logs and user questions can further refine prompts, guardrails, and dataset documentation, creating a city-specific knowledge asset that’s hard to replicate quickly.
Hybrid
Vector Search
Medium (Integration logic)
Context window cost and retrieval quality as the volume and diversity of open datasets, metadata, and documentation grows; also governance and privacy constraints around which internal vs external datasets can safely be exposed through natural-language interfaces.
Early Adopters
Unlike generic chat-with-your-data tools, this implementation is tailored to a single city’s open data ecosystem, policy context, and governance rules. The differentiator is not the core model but the deep integration with DC’s data catalog, metadata, and public engagement workflows, making it more useful and safer for civic use than off-the-shelf AI chatbots.