The second day of Bentley’s Year in Infrastructure Awards featured an engaging panel of veteran jurors reflecting on what makes innovation last. Innovation, they agreed, is “sticky” only when it becomes outcome-focused and embedded in daily operations rather than treated as a one-off. One panelist framed it plainly: you can’t just say it’s one and done. You have to look at long-term objectives and how new tools connect to everything else you’re already using.
The conversation turned to technology as a great equalizer. AI and natural language interaction are opening the industry to a more diverse workforce and breaking down silos between disciplines. One juror recalled his doctoral work on convergence technology in the late 1990s. The difference today, he said, is pace. Changes that once took a century now occur in a fraction of the time.
The jurors also noted a shift in how projects are executed. Technology enables systems to share information seamlessly. Engineers and contractors can now collaborate closely, catching mistakes immediately instead of discovering them as costly errors later.
A powerful example came from structural engineer Steven Baldridge, who analyzed the Seaside collapse from his office in Honolulu in a single night. What would have taken months was accomplished in roughly six hours. The panel framed this as a turning point: technology is enabling teams to move from reactive analysis after a failure to prescriptive action before one.
The theme of prevention over reaction set the stage for the undercurrent throughout the entire program: climate. As infrastructure ages and climate disasters intensify, the industry faces a world where extreme weather, shifting rainfall, and geological instability are no longer anomalies but operating conditions. By the afternoon, the Geospatial and Reality Modeling award category would offer the most compelling answer yet.
Geospatial and Reality Modeling: Three Projects, One Winner
I had the honor to participate as a judge in this category, and to witness the incredible work that many organizations and people are completing around the world. Three projects from three continents demonstrated how reality capture, AI, and digital twins are reshaping how infrastructure is understood, monitored, and increasingly defended against climate-driven threats. All three finalists were a testament to the sticky innovation the jurors had described earlier.
Victoria’s Transport Reality Capture (Australia)
Victoria’s Department of Transport and Planning built a single cloud-hosted repository on Bentley Infrastructure Cloud to consolidate scattered reality capture data across the agency. The AUD 5 million project, completed in February 2025, reported a 50% reduction in data discovery time and a 20% uplift in data reuse. The platform also uses AI to extract road assets from panoramic images.
Drone Inspections at the Port of Dover (UK)
The second finalist brought drone-derived 3D models to one of the world’s busiest ferry ports. Haskoning inspected 59 assets, capturing over 3,520 defect records through a modular pipeline. The £444,940 project reduced inspection hours by 33%, delivered approximately £150,000 in ROI, and eliminated 60 hazardous person-hours. Drone-based models reduce or eliminate traditional risks like rope access and confined-space entry while producing a richer record than visual inspection alone.
GeoShield: AI-Driven Landslide Prediction (Philippines)
The winning entry was GeoShield, an AI-driven landslide risk platform for a hydropower plant on the island of Mindanao, and the project that most directly answered the program’s challenge to move from reaction to prevention. The Philippines is among the most disaster-prone countries globally, and rising rainfall intensity, more frequent typhoons, and shifting soil moisture patterns are making slope instability worse, putting hydropower infrastructure, downstream communities, and the energy grid at growing risk.
The roughly $200,000 pilot combined lidar, photogrammetry, electrical resistivity tomography, GIS data, and field validation into an iTwin digital twin. It predicted two landslide-prone areas before failure, allowing mitigation in advance, and reports 85% prediction accuracy and 91% simulation accuracy. A comparable landslide at another Mindanao facility cost approximately $27 million; GeoShield is credited with roughly $6.8 million in avoided losses.
Its significance extends beyond ROI. GeoShield represents a new category of climate-resilience infrastructure: digital systems designed to anticipate disasters rather than respond to them. Its digital twin ingests real-time geospatial and geophysical data to model changing conditions continuously, rather than relying on static assumptions from a pre-climate-change era. It offers a template for how infrastructure owners worldwide can adapt risk management to a planet that no longer behaves the way it did when those systems were first designed.
Looking Ahead
The prevention theme carried into the panel’s look at the future. When asked what the industry would be discussing at YII in 2031, presenters converged on physical AI, spatial computing, and robotics on job sites. Safety transformation was another theme, using digital twins to rehearse dangerous operations before anyone steps on site. The underlying message was consistent: regardless of what the technology is called, what matters is sharing the right information to make the best decisions. As one panelist put it, if you live in a 2D world and want to use AI, it’s going to be very tough. We naturally live in a 3D world.
Day 2 came down to a simple contrast: an engineer in Honolulu analyzing a collapse in hours instead of months, and a digital twin in Mindanao flagging landslides before they happened. The Geospatial and Reality Modeling finalists demonstrated that the gap between what’s possible and what’s standard practice is narrowing fast, and GeoShield’s victory underscored that the highest-impact applications of geospatial technology are the ones that prevent disaster rather than respond to it. The challenge for 2031 and beyond is clear: we must build digital infrastructure that can keep pace with a changing physical one.
