Field BIM: Taking the Model to the Construction Site
A practical field BIM guide for coordinators: field-ready models, mobile viewers, model-based layout, issue capture, and verification on site.
Go deeper with Archgyan Academy
Structured BIM and Revit learning paths for architects and students.
The model usually dies at the site gate
Most projects invest months in a coordinated model. Disciplines share, clashes get resolved, the federated model is clean, and everyone signs off. Then construction starts, and the site team prints PDFs. The model that cost thousands of coordination hours sits on a server while the foreman works from a drawing that was exported three weeks ago.
This is the field BIM gap, and it is where a large share of BIM’s promised value quietly evaporates. Rework on site rarely comes from a bad model. It comes from the site not using the model: an outdated drawing, a dimension scaled off a print, a services route that was re-coordinated after the subcontractor’s set was issued.
Field BIM closes that gap. It is the set of workflows that put the current, coordinated model in the hands of the people building the thing: viewers on tablets, layout instruments driven by model coordinates, issues captured against model elements, and installed work verified against design geometry. This guide walks through each piece from the perspective of the BIM coordinator or manager who has to make it work.
What field BIM actually covers
Field BIM is not one tool. It is four connected workflows, and it helps to name them separately because they have different owners, different tooling, and different failure modes:
Model access on site. Viewing the federated model, drawings, and element data on a tablet or phone, in the container where the work happens. This is the foundation; nothing else works without it.
Model-based layout (setting out). Taking points from the model and staking them physically with a robotic total station or GNSS rover, instead of pulling tapes from gridlines on a paper plan.
Field data capture. Raising issues, quality inspections, and safety observations against model elements and locations, so problems flow back to the coordination process instead of dying in a phone’s photo gallery.
Verification. Comparing what was actually installed against what was designed, using 360-degree photo capture or laser scans overlaid on the model.
A project can adopt these incrementally. Model access alone already changes site behavior. Layout and verification need investment in hardware and skills, and most teams add them once viewing is routine.
Why drawings-only on site wastes the model
It is worth being precise about what goes wrong when the site works purely from prints, because these are the arguments you will need when someone asks why the tablets are worth it.
Version lag. A printed drawing is a snapshot. Between print and installation, coordination continues. The classic failure: MEP re-routes a duct after a clash, the model and the current drawing revision are updated, but the sub is installing from the set in the site office. With a connected viewer, the current published model is the only thing anyone can open.
Lost information. A drawing shows a fraction of what the model knows. The model carries the duct’s system, its insulation spec, its hanger requirements, the fire rating of the wall it penetrates. On paper, the installer sees a line and a tag, and either goes looking through specs or guesses.
No spatial reasoning. Congested ceiling voids, risers, and plant rooms are exactly where 2D fails. A section every few meters cannot communicate a three-trade stack sequence. Rotating the actual coordinated geometry on a tablet in front of the installer settles in seconds what a drawing discussion drags out for minutes.
One-way communication. Paper cannot report back. Every issue found on site has to be re-described in an email or photographed with no location context. The office then spends time just figuring out where the problem is.
None of this means drawings disappear. Contractual deliverables remain drawings on most projects, and a laminated plan still beats a tablet in the rain. Field BIM adds the model alongside the drawing set and makes the model the source both derive from.
What “field-ready” means for a model
The most common reason field BIM adoption dies is not the software. It is that the model handed to the site was never prepared for it. A coordination model optimized for clash detection is not automatically usable by a foreman with two minutes and cold hands. Field-ready means:
Published, not work-in-progress. Site users must only ever see information from the Shared or Published state of your common data environment, never someone’s live working file. Wire your field viewer to the published area of the CDE and nothing else.
Federated and filtered. One combined model, but with sensible default views: per-level, per-zone, per-trade. A drywall crew does not need structural rebar visible. Prepare saved viewpoints or views per trade and level so nobody starts from a full-building soup of geometry.
Named for humans. On site, elements get found by search. “L03-CorridorC-FCU-07” is findable; “Fan Coil Unit (2)” is not. This is where your naming convention proves its worth, or its absence gets expensive.
Light enough to load. Tablets choke on what workstations tolerate. Purge unused content, drop over-modeled manufacturer geometry, and export to the viewer’s optimized format (NWD, or the platform’s own streaming format) rather than raw authoring files.
Carrying the data the field needs. Installers care about a short list of parameters: system, size, level, zone, spec reference, install status. Make sure those survive the export. A perfect geometry export with stripped parameters is a 3D picture, not a model.
A practical test before rollout: hand the tablet to a site engineer who was not in the pilot and ask them to find one specific valve and tell you its system and spec. If that takes more than two minutes, the model is not field-ready yet.
Choosing the mobile viewer stack
The viewer market is mature, and the honest answer is that the leading tools overlap heavily. The bigger decision is usually made for you: if the project CDE is Autodesk Construction Cloud, ACC Build is the path of least resistance; on Dalux Box projects, Dalux Field is. A quick orientation:
| Tool | Strengths | Typical fit |
|---|---|---|
| Dalux Field | Very fast free viewer, strong drawing-model overlay, simple for trades | Europe-heavy projects, main contractors, large trade rollouts |
| Autodesk Construction Cloud (Build) | Deep integration with ACC CDE, issues, RFIs, forms in one place | Projects already running ACC as the CDE |
| Revizto | Strong issue tracking shared between office and field, good federation | Coordination-led teams wanting one issue platform everywhere |
| BIMcollab GO | Clean BCF-native issue flow, lightweight viewer | openBIM teams already using BIMcollab for coordination |
| Trimble Connect | Ties into Trimble layout hardware and Tekla models | Structures, steel, and layout-heavy scopes |
| Procore (Models) | Field management platform first, models added alongside | US-market projects run on Procore |
Selection criteria that actually matter in practice: does it work offline (site connectivity is never as good as promised), can it overlay 2D drawings on the model (trades trust drawings first, and the overlay builds trust in the model), does it round-trip issues in an open format like BCF, and can a subcontractor use it without a license negotiation. Run a four-week pilot with one trade on one level before you commit the project.
Model-based layout with robotic total stations
Setting out is the field workflow with the most direct, measurable payoff. The traditional method - a site engineer with a plan, a tape, and gridline offsets - is slow and reproduces whatever errors live in the print. Model-based layout inverts it: layout points are placed in the model in the office, pushed to a robotic total station or GNSS rover, and staked on site by one person following a prism.
The workflow, end to end:
- Place points in the model. Sleeve centers, hanger inserts, wall lines, anchor bolts, MEP penetrations. Tools like Autodesk Point Layout or Trimble’s field points add these as model objects with IDs.
- Check the coordinate basis. The layout instrument works in the site survey coordinate system. Your export must use shared coordinates that reconcile with the survey control network. This is the single most dangerous step; a model exported on internal project coordinates will stake points in a field somewhere.
- Push to the instrument. Export the point set to the layout software (Trimble FieldLink, Leica iCON, Topcon), synced with point IDs matching the model.
- Stake and record. The instrument guides the operator to each point. Crucially, it also records as-staked positions, giving you a quality record for free.
The payoff numbers convince people: a two-person crew laying out a few dozen points a day with tape and plan versus one operator staking hundreds of model-driven points, with each point traceable to a model element. For penetrations and inserts on concrete pours, where a missed sleeve means core drilling, the avoided cost is vivid enough that contractors rarely go back after the first project.
For the coordinator, the takeaway is upstream: layout quality is decided by model quality. If hangers and sleeves are not modeled, or are modeled approximately, model-based layout has nothing trustworthy to stake.
Issue capture: closing the loop back to coordination
Site walks find problems. The question is what happens next. The failure mode everyone recognizes: a photo on someone’s phone, a verbal mention in the weekly meeting, no location, no owner, no closure. Field BIM replaces that with the same issue lifecycle you already run in coordination.
The mechanics: the engineer on site opens the model or drawing at their location, taps the element or spot, and raises an issue with a photo, a description, an assignee, and a due date. Because the issue is anchored to a model element and location, the office sees exactly where it is. If your platform speaks BCF, the issue flows into the same coordination environment where clash issues live, and gets triaged in the same meetings, by the same rules.
Two disciplines make this work. First, make the site raise issues in the tool, not around it. That happens only if raising an issue is faster than sending a WhatsApp photo, which is a setup and training problem, not a willpower problem. Second, close the loop visibly. When issues raised from site get answered within a day or two, the site keeps raising them. When they vanish into a backlog, the site goes back to WhatsApp within a fortnight.
Verification: comparing built to designed
The last workflow answers a question coordination cannot: was it actually built the way we coordinated it?
360-degree photo capture is the accessible tier. A site engineer walks the floor with a 360 camera on a helmet or pole; the platform (OpenSpace, or the capture tools inside the major CDEs) pins the imagery to the floor plan and, at the stronger tier, side-by-side with the model. Weekly walks give you a dated visual record of every corridor. This catches the obvious: work in the wrong place, sequence deviations, and it settles disputes about what was visible before a wall closed up.
Laser scan verification is the precision tier. Scanning installed work and overlaying the point cloud on the coordination model shows deviation numerically: slab flatness, riser positions, whether the in-ceiling services match coordinated routes before the ceiling closes. On projects with prefabricated assemblies arriving to tight tolerances, scan verification before and after critical installs is what keeps the prefab promise honest.
The verification cadence that earns its cost on most building projects: 360 capture weekly on active floors, scans at closing-up milestones (before ceilings, before shaft closures, after structure per level). Verify at the moment the work is about to become inaccessible, because that is the last cheap moment to fix it.
The as-built feedback loop
Everything the field learns should eventually flow back into the model, because the model outlives construction. The as-staked layout points, the verified scan deviations, the accepted substitutions raised through field issues: on a project heading toward a structured handover, these are the difference between an as-built model that reflects the building and a design model relabeled “as-built” on the title block.
Practically, set a rule early: any field deviation above an agreed tolerance, and any substitution, must result in either a model update or a documented waiver. Assign that update work explicitly (usually to the originating discipline’s modeling team, tracked through the same issue system). Projects that leave “update the model” as an implied good intention hand over fiction.
Who owns field BIM
Field BIM sits exactly on the seam between the BIM team and site management, which is why it falls on the floor when nobody names an owner. A split that works on most projects:
- BIM coordinator / manager: model readiness (federation, views, naming, exports), platform administration, coordinate basis for layout exports, and the BCF bridge back to coordination.
- Site engineering: layout execution, capture walks, issue raising, and first-line triage of field issues.
- Trade contractors: consuming the model for their scope, raising issues on their work faces, and providing install-status updates where the project tracks them.
The role this creates on larger sites, sometimes titled field BIM engineer or site BIM coordinator, is one of the better entry points into BIM careers right now: it needs someone who understands both the model and the mud, and there are far fewer of them than the market wants.
Common mistakes
Rolling out to everyone at once. A big-bang rollout to twelve trades generates twelve flavors of confusion and one bad reputation. Pilot with one trade, one level, one workflow. Expand on demonstrated wins.
Handing the site the coordination model raw. Full-detail federated models with no trade views, no viewpoints, and coordination-oriented naming convince the site that “the model is too complicated.” The model was fine; the preparation was skipped.
Ignoring offline reality. If the viewer needs live connectivity and the basement has none, the tool is dead in exactly the congested areas where it is most needed. Test offline sync before rollout, not after.
Layout from unreconciled coordinates. Exporting layout points without verifying shared coordinates against the survey control network. This one physically moves buildings. Verify with known control points on every export until the pipeline is proven.
Treating field issues as a separate world. Field-raised issues managed in a different tool, by different rules, than coordination issues. You end up with two truths. Route everything through one lifecycle, BCF where possible.
Buying hardware before workflows. Tablets and total stations purchased in a burst of enthusiasm, then used as expensive PDF readers because nobody defined the model-preparation and data-flow steps that make them useful.
No visible response to field input. The fastest way to kill adoption is to let site-raised issues go unanswered. The site gives the office exactly one chance to prove the loop works.
How to start
- Start with viewing on one active project. Stand up the mobile viewer your CDE already includes, prepare per-trade views for one level, and give tablets to two site engineers and one willing trade. Four weeks, then review.
- Make field-readiness part of your model QA. Add the field checks (naming searchable, views prepared, export size, parameters surviving) to the same pre-share checklist you already run for coordination.
- Wire issues to BCF from day one. Even in the pilot, route field issues into the coordination issue register. The loop is the product; the viewer is just the door.
- Add layout where the pain is worst. Penetrations and inserts on the next concrete package is the classic first layout scope. Partner with the survey team on the coordinate basis before anything is exported.
- Schedule verification at closing-up milestones. Start with 360 capture weekly; add scans where tolerance risk justifies them.
- Name the owner. Someone on the project must own field BIM by name, with model preparation and platform admin in their actual job description, not as a side quest.
Field BIM is where coordination stops being a report and starts being a building. The skills it demands, model preparation, CDE discipline, issue management, coordinate literacy, are the working core of the BIM coordinator role. If you are building toward that role, our BIM and Revit courses teach these workflows the way firms actually run them, taught by a working BIM Coordinator.
Level up your skills
Ready to learn hands-on?
- Project-based Revit & BIM courses for architects
- Go from beginner to confident professional
- Video lessons you can follow at your own pace