By Alec Pestov, CEO, vGIS Post-event perspective from the Common Ground Alliance Conference and Expo 2026

At CGA 2026, the conversations felt familiar, but more urgent.

Damage prevention is no longer a knowledge problem. The industry knows what needs to be done. The challenge now is execution, and the data infrastructure that supports it.

Despite years of measurable progress in processes, tools, and awareness, damage continues to occur at scale. The Common Ground Alliance’s most recent DIRT Report documented approximately 190,000 damages to underground infrastructure across North America in a single year, with an estimated economic impact exceeding $30 billion annually when factoring in repair costs, service disruptions, and third-party liability. That figure reinforces a sobering reality: this is not a gap in intent. It is a gap in consistency.

That tension shaped much of the discussion throughout the event.

Where the System Breaks Down

If there was one issue that surfaced repeatedly, it was record sharing.

Everyone agrees it is critical. It underpins planning, coordination, and safe execution at every stage of a project. But in practice, it remains deeply inconsistent. According to CGA’s own research, inadequate or inaccurate locating continues to rank among the top root causes of excavation damage, accounting for a significant share of preventable incidents year over year.

Information is often fragmented across systems, formats, and stakeholders, and it rarely flows in a way that supports real-time decision-making. By the time a crew arrives on site, the data that should guide them is frequently incomplete, outdated, or inaccessible in the field.

CGA data illustrates the downstream effect clearly. Even when excavators follow the correct notification process, they face roughly a 38% probability of being unable to begin work as scheduled due to incomplete or delayed locate responses. The process was followed. The system did not fully support the outcome.

This is not a new finding. The American Society of Civil Engineers has consistently highlighted aging underground infrastructure and insufficient as-built records as compounding factors in excavation risk, noting that a significant portion of buried utility data in North America is either incomplete or has not been updated to reflect decades of construction activity. When records are missing or unreliable, even the most diligent crews are working with incomplete information.

AI Is Advancing, But Built on the Same Foundations

Artificial intelligence was a prominent theme across CGA 2026, with clear and growing interest in its potential to improve damage prevention outcomes through predictive analytics, automated risk scoring, and smarter resource allocation.

But the conversations were notably grounded.

AI is not a shortcut. Its effectiveness is directly tied to the quality, completeness, and accessibility of the data it relies on. Predictive models trained on fragmented or inconsistent records will not produce reliable predictions. Computer vision systems applied to incomplete subsurface data will not improve field accuracy.

Research from McKinsey and others has consistently noted that data quality and integration remain the primary barriers to realising AI’s value in infrastructure-intensive industries, not the algorithms themselves. The industry is beginning to internalise that message.

As adoption grows, the clearer it becomes: meaningful progress with AI in damage prevention depends on first strengthening the data foundations it requires. Better models applied to incomplete records will not produce better outcomes.

The Shift Is Happening Before the Dig

What stood out most at CGA 2026 was a meaningful shift in where the industry is focusing its attention.

Less emphasis on reacting to damage after it occurs. More emphasis on what happens before crews arrive, at the planning and pre-excavation stage, where the key assumptions that drive field behaviour are formed.

Because once excavation begins, uncertainty becomes risk. And risk in this context carries real consequences: service outages affecting hospitals, transit systems, and emergency infrastructure; worker injuries; and liability costs that can run into the millions for a single incident.

Improving visibility before the dig leads to more confident planning, fewer surprises during execution, and a meaningful reduction in the last-minute field judgements that contribute disproportionately to damage events.

The DIRT Report supports this framing. A small number of recurring root causes, including inadequate locating, failure to maintain clearance, and excavation outside the tolerance zone, continue to drive the majority of damages. Addressing those causes requires earlier and more reliable access to subsurface information, not simply more reminders to be careful.

At vGIS, this aligns directly with what we are seeing across projects. Utility owners and municipalities are increasingly focused on making subsurface information more accessible and actionable in real-world field conditions, from LiDAR-based reality capture during open-trench work to centimetre-accurate augmented reality overlays that persist long after backfill.

The Path Forward Is Clear, But Not Yet Consistent

CGA 2026 did not introduce new problems. It reinforced familiar ones, with greater clarity about where the leverage points lie.

Better record sharing. Stronger data foundations. More practical, field-ready technology that works in the hands of the people who need it most.

The CGA DIRT data points to a clear conclusion: the industry is not lacking solutions. It is struggling to apply them consistently at scale, across a fragmented landscape of stakeholders, systems, and workflows that rarely connect cleanly.

That is the real challenge in front of the industry.

Final Thought

CGA 2026 did not highlight a lack of solutions. It highlighted the need to connect them, across systems, across stakeholders, and earlier in the lifecycle of every project.

Because progress in damage prevention will ultimately come from improving how information flows, not just how it is created. The technology to do that exists. The question now is whether the industry can deploy it consistently enough to move the number that matters most.

Nearly 190,000 damages a year is not a knowledge problem. It is an execution problem. And execution starts with data you can trust, in the hands of the people who need it, before the first shovel breaks ground.