Transit Insight Agent: An AI assistant for complex transit analytics

The Problem

Data is heavily constrained by a "silo, context and rigidity" gap that compromises operational intelligence. Currently, fragmented data silos prevent analysts from combining critical metrics—such as linking service levels directly to trip-level ridership—into a cohesive performance narrative. This fragmentation is compounded by severe labour inefficiencies, forcing staff to dedicate extensive manual effort to raw data preparation, extraction and alignment instead of strategic analysis. Furthermore, existing business intelligence dashboards are rigidly pre-aggregated and incapable of adapting to complex, ad-hoc inquiries.

The Solution

The Transit Insight Agent, an intelligent data assistant designed to democratize and automate complex transit analytics, has been developed to bridge this gap. The system integrates fragmented data pipelines—such as routing, scheduling and automated passenger counter (APC) data—into a unified semantic layer.

By leveraging an advanced AI agent interface, the platform automates intensive data prep labour of extracting, joining and aligning disparate datasets. Moving beyond the limitations of rigid, pre-aggregated business intelligence dashboards, the Transit Insight Agent interprets natural language questions to dynamically generate ad-hoc, transit analytic insights. This enables transit planners and operational staff to instantly answer complex, long-tail inquiries through an automated conversational interface, shifting their focus from manual SQL and Excel work to proactive, data-driven strategy. This tool was built using a combination of open source tools with gcp infrastructure (e.g. Gemini, Bigquery costs).

The Impact

By introducing an intuitive AI interface that interprets natural language, the platform bypasses rigid, pre-aggregated dashboards to provide instant, transit analytics insights for complex, ad-hoc inquiries. This operational shift eliminates analytics bottlenecks, allows transit staff to focus entirely on high-value strategy, and delivers a holistic view of performance to optimize the rider experience.

Lessons Learned

Currently in development -- not yet operational

Launch year: 2026