Pulse Logic is a data‑driven architectural construct that uses live and recorded mobility streams—buses, tube, bicycles and pedestrian flows—to interrogate how infrastructure manages density, proximity and control. So far, AI has mainly been used for person detection and tracking, but the project is now moving toward custom models that read patterns over time, diagnose where the network is consistently under stress, and flag spatial “problem zones” in the city. These models will not only detect where congestion and stationary crowds occur, but classify different types of waiting, risk and discomfort, effectively turning AI into a diagnostic tool that produces design briefs for intervention. As an architectural project, Pulse Logic will translate these diagnostics into concrete spatial proposals: reconfigured concourses and platforms, new upper‑ground waiting chambers, diversions and public spaces that are specifically shaped by the detected patterns of use. The construct itself becomes a piece of architecture too—an interactive control room/interface where rules, thresholds and routes can be edited and tested against live or simulated data. Ultimately, Pulse Logic aims to move from simply visualising flows to actively re‑writing the spatial logic of transit nodes, using AI not only to see the city but to propose how and where it should be rebuilt.