We've previously written about why our underground infrastructure keeps getting struck – the nutshell is that thousands of subsurface lines have inaccurate or outdated locational data and a disconnect between what's actually underground and what's on record leads to property damage, injuries, and cost overruns. The workflow Esri and Pix4D announced in February 2026 closes that gap at the exact moment it happens: just before a trench gets backfilled. It's not groundbreaking news, but it's an extremely useful workflow that hasn't received as much attention as it deserves, given how directly it addresses a problem we've covered ourselves.
What the Phone-Based AR Workflow Really Does
It pairs Pix4D's PIX4Dcatch mobile app with an RTK device to produce georeferenced, high-precision 3D scans of trenches and infrastructure right from a cell phone. That data then publishes directly to Esri's ArcGIS Online platform, where it manifests as a 3D scene layer comprised of viewable, 3D point clouds. Compared to the conventional method of measuring depth and offset in the field with a paper or simple form, then digitizing that document in a GIS later, this is a significantly different workflow with a significantly higher likelihood of accuracy (no typos, units transposed, or records lost in the trench-backfilling process).
How This Stacks up to a Traditional Utility Locate and Record Process
The standard method of documenting an exposed utility span is for a field technician to measure depth and offset by hand, record that data on a paper or tablet form, then have someone else later enter that information into a municipal or utility GIS. Every time that handoff happens, the potential exists for human error: a number gets read wrong, a number gets transposed, an individual record gets skipped entirely. This phone-based RTK+AR process reduces several of those dependencies: the capture itself produces a georeferenced point cloud (not just coordinate data) that can publish directly into the geodatabase without an entire manual data transfer stage.
The Extra Step That Actually Matters: AR Verification Before Closing the Trench
This is the piece that most directly relates to the problem of the documentation gap. This process enables visualizing existing design data in augmented reality on the job site, allowing a field crew to compare as-designed to as-built in real time before the trench gets backfilled. Once a trench is backfilled, verifying the exact location of utility installations requires either trusting a record or excavating again. Spotting placement conflicts while the trench remains open turns what could have been a future utility hit into a quick five-minute correction on site.
Why We Care About This More Than Another Software Integration
Software companies unveil hundreds of integrations a day, and most of those integrations are extremely minor. We're highlighting this one specifically because it reaches directly into the mechanism that creates the underground utilities problem in the first place: not a lack of GIS tools in general, but a lack of reliable, up-to-date records of what was actually put in the ground, and where it was placed, that any given excavation crew can consult before starting to dig. Pix4D frames it as turning "hidden infrastructure into usable data" and that is the very gap in our own description of the underground utility problem.
Frequently Asked Questions
Does this replace a utility locate survey?
No. This is a document-and-verify workflow for utilities that have been located; it is not a detect-and-expose workflow for finding unlocated buried utilities.
What level of accuracy can it feasibly support?
That depends more on the RTK device being used, not the capture software itself. Confirm with your vendor for the field hardware you have in mind.
Can it work for existing, unmapped lines as well as new installations?
Yes. For new construction, this prevents the documentation gap entirely. For systems that already exist, it can document the next time an existing line happens to be located.
What should I be on the lookout for if this kind of workflow is the avenue I'm considering for AR-based utility documentation?
RTK compatibility – the overall accuracy you're going to be able to get depends as much on the hardware as the capture app itself. A single large site still requires a drone or traditional ground survey for broad coverage; this is not a solution for examining an entire property or worksite, only a single, small trench. Focus on the end result, not just the raw data: knowing how seamlessly that scan can be fed back into your team's current geodatabase to be referenced on future excavations. Creating an AR tool that can be used to compare as-designed to as-built after-the-fact only goes so far if field crews can't recognize the significance of a discrepancy. Training matters.
The Larger Pattern to Follow
This fits into a larger theme we've been noticing lately here on this blog: the tools of the geospatial industry are becoming increasingly designed to bridge the gap between data capture and data use, and not merely capture more data. An AR model sitting in an app, unintegrated into the GIS platform your team actually uses on-site, isn't going to fix the underground utilities problem. A model that directly and automatically feeds into your GIS platform will.
