From impossible to in-demand

When aging infrastructure meets shrinking budgets, the result is usually a slowdown for traffic, teams and technology. But for Alex Garcia and his team at GeoVerra, one of Canada’s largest geomatics firms, it was a signal to hit fast-forward.

By training artificial intelligence to inventory entire corridors in hours rather than weeks, GeoVerra is transforming how the public and private sectors manage the built environment. As the national manager of mobile solutions, Garcia is focused on a singular goal: integrating mobile mapping with AI to identify, extract and manage assets with unprecedented speed.

As a rule, Garcia embraces geospatial technology that solves problems faster and more efficiently. By utilizing AI models within Trimble Business Center (TBC), GeoVerra can automatically recognize everything from railway signals to highway signs, meeting a desperate need for accelerated project delivery.

Precision at scale

Asset management has shifted from a back-office function to a strategic priority. To spend taxpayer money effectively, municipalities must know exactly what they own and what condition it is in.

GeoVerra, which employs more than 500 people across North America, uses Trimble MX90, MX50 and MX9 mobile mapping systems to capture high-resolution imagery and point clouds at highway speeds. This eliminates the need for lane closures and, more importantly, keeps crews out of hazardous environments.

The real magic happens during processing. Traditional manual extraction for a project covering hundreds of kilometers could take months — a timeline that often makes such projects unfeasible. AI-driven extraction changes that math. While TBC includes pre-trained models for common items like poles, GeoVerra’s edge lies in customization.

Because a utility pole in Ontario differs from one in British Columbia, GeoVerra trains custom models by providing the AI with representative samples. Once trained, these models are used repeatedly, making each subsequent project more cost-effective.

Real-world velocity

The results are staggering. During a rail project in the Greater Toronto Area, GeoVerra collected data for 970 kilometers of corridor. A trained AI model extracted 33 different asset types at a rate of 15 kilometers of track in just 25 minutes.

A similar story unfolded on an Ontario highway project. Along 125 kilometers of road, the team extracted more than 8,000 individual features across 21 categories. Because these assets remain consistent across the province, GeoVerra can now bid on new contracts knowing their established models are ready to work on Day 1.

Competitive edge

By consolidating these workflows into the TBC platform, GeoVerra has eliminated the need for expensive, disparate software licenses and messy data transfers. Garcia also credits his team’s relationship with Trimble’s developers, valuing "the direct access to technical staff who supported GeoVerra's early adoption of these tools."

As interest in asset lifecycle management grows, GeoVerra is eyeing even larger opportunities, from citywide inventories to airport pavement analysis. By turning "impossible" datasets into actionable insights, this team isn't just mapping the world — it’s helping to keep it running.