SmartBuild · Capabilities

Evidence at source, validated as the work happens

SmartBuild is a construction quality and compliance platform. Instead of reconstructing what was built from an office, it captures the proof at the action: pinpoint location and depth, camera and computer-vision checks, a workflow that locks the sequence, and full offline operation. The result is evidence of the build, not a layer drawn after the fact.

Get a demo →

The frame

Proof of the build, not a layer drawn later

On most rollouts the record is assembled after the fact. Contractors rush at month-end and cut corners on depth, compaction and documentation. Photos get reused across stretches, test reports are fabricated, and the data is corrupted at every handoff up the hierarchy from field crew to site engineer to contractor to PM to client. Audits surface the problem months later, when remediation costs many times what prevention would have.

SmartBuild inverts that. It embeds compliance in every field action, so the evidence is captured at source, the moment the work happens. The map is the interface, but the product is operations-first: it starts from the quality and compliance of the work, not from a GIS layer.

The capture pillars

Four things SmartBuild does at the action

  1. 01

    Pinpoint location and depth, captured in workflow

    Sub-5cm captured-point accuracy, CORS-corrected, with NMEA decode and PPK support. Position and depth are recorded together at the moment of construction, as an enforced step, not a separate as-built pass.

  2. 02

    Auto-fill from the camera, so there is nothing to fabricate

    The camera module runs blur, size, lighting and stability checks and can auto-capture. Computer vision reads values straight off the image, which extracts the value rather than asking a crew to type it. Auto-fill is anti-tamper by design.

  3. 03

    The next step opens only after the prior evidence validates

    Workflows are customizable sequences of entities. The sequence is locked: a crew cannot move forward until the current evidence passes its checks, so the compliance is a property of the work rather than a form filled in afterwards.

  4. 04

    Full offline operation, with no data loss

    Maps, layers and workflows download to the device. Capture saves locally and sync is deferred to end of day. The workflow stays locked and enforced even with no signal, so the field never has to choose between coverage and compliance.

Inside an entity capture

One captured entity, four kinds of evidence

The unit SmartBuild captures is an entity: a trench span, a manhole, a joint closure, a pole, a restoration. Each entity can carry image, video, form and map-mark evidence, in any order, with any element skippable, and counts, durations and forms configured per workflow. That single, flexible record is what resolves into proof of the build.

  1. 01

    Image, photographic evidence checked on capture

    The image carries the visual proof and feeds the computer-vision checks. Quality gates on blur, size, lighting and stability mean the evidence is usable, not a blurry afterthought.

  2. 02

    Video, a moving record where a still is not enough

    Video captures sequences a single frame cannot, for entities where the process matters as much as the end state.

  3. 03

    Form, structured values auto-filled where possible

    Forms hold the structured data for the entity. Where computer vision can read a value from the image, it auto-fills the field, so the speed comes from the camera and the numbers are not hand-typed.

  4. 04

    Map-mark, the pinpoint position on the route

    The map-mark fixes where the entity sits, captured at sub-5cm with CORS corrections, so the record knows exactly where on the route the work was done.

Accuracy and cadence

Captured once, at sub-5cm, every nine metres

These are technical performance targets for how SmartBuild captures the build, not customer outcomes. They describe the bar the platform is built to: position and depth fixed precisely, sampled densely along the route, so the record is detailed enough to act on.

Precise records also change what happens later. When a fault has to be found, an accurate as-built can point to a single dig instead of multiple exploratory pits.

One validated record, formed at the point of work.

As the capture point moves along the trench, it drops position-and-depth markers and gathers image, video, form and map-mark evidence. They resolve into a single validated record, fixed to the route, ready to feed the digital twin.

Why captured this way

Compliance becomes a property of the work

Capturing at source, in a locked workflow, with auto-fill from the camera, closes the gaps that the old way leaves open. Position and depth are fixed precisely as an enforced step. Values are read off the image instead of typed, so there is nothing to fabricate. The sequence cannot skip a check even with no signal. What you are left with is a record that proves the build, captured at the action rather than reconstructed afterwards.

< 5cmcaptured-point accuracy, CORS-corrected, with NMEA and PPK (technical target)
9mdepth captured along the route, configurable per workflow
6computer-vision use cases proven as POCs to date

Field validation

Geo-stamped evidence, checked on capture

At each entity, the field app presents a checklist of compliance steps: trench depth, duct integrity, backfill condition, as-built match. Each step opens only when the prior evidence is validated, and each photograph is geo-stamped as it is taken, so the evidence carries the place it came from.

Nothing waits on a signal. Capture saves to the device and the record is queued, then syncs to the web console at end of day, which is why the workflow can stay locked and enforced in the field.

What capture-at-source makes possible

Evidence that holds up

Position, depth and condition are captured at the action, in a locked workflow, so the record reflects what was actually built rather than what was reported later.

Speed without typing

The camera does the data entry. Computer vision reads the value from the image and auto-fills the form, which is faster for the crew and removes the chance to fabricate a number.

A record the twin can trust

SmartBuild is designed to feed the NetworkAccess digital twin through SmartInventory integration, on the v1 roadmap, so the as-built becomes the network of record.

Status

Built device-agnostic, proven in computer-vision POCs

Capture is device-agnostic. SmartBuild connects to CORS, applies corrections to the chip and decodes NMEA directly, with PPK support and no SDK lock-in, so it is not tied to one receiver. SmartMeasure is available as a higher-grade GNSS receiver where a project wants one, but the story does not depend on it. SmartBuild v1 is in progress, and six computer-vision use cases are already proven as POCs, with more in the pipeline as roadmap items.

Proven to date 6 computer-vision use cases proven as POCs, with further use cases on the roadmap
  • Trench
  • Duct
  • Manhole
  • Last-mile cable
  • Manhole restoration
  • Pole inspection
Get started

See SmartBuild capture the build.

Walk through the field app and the web review console with our team, on a workflow that mirrors your own construction sequence.

Get a demo →