Digital twins, virtual models mirroring real-world physical systems in real-time, are transforming how we design, simulate, and operate complex environments before real‑world deployment. Building on our work at the Digital Twin Cities Centre (DTCC), I will give an overview of projects such as automatic urban geometry and mesh generation, digital control rooms for construction site management, high‑fidelity simulations of wind, pollution, and noise, and the DTCC data platform that underpins AI‑driven twins across an entire city.
 I will also discuss how AI can accelerate and enable each stage of this pipeline: speeding up research exploration and systematic literature studies; co‑piloting programming and, where appropriate, generating library code and web interfaces; constructing city‑scale 3D models from heterogeneous geospatial data; and learning surrogate models (e.g., neural operators and reduced‑order surrogates). Finally, I will reflect on how stronger AI tools shift the bottleneck from implementation capacity to attention. When researchers are amplified, the decisive skill becomes problem selection. I’ll propose a pragmatic triage—prioritizing by Impact, Tractability, and Data Readiness—to focus effort where it changes practice rather than merely what is easy to do.