The maintenance execution problem

A preventive maintenance procedure for a complex industrial asset can run 50 to 200 discrete steps. The technician consults the procedure document, looks up at the equipment, performs the step, looks back down at the document, confirms, advances to the next step. Each transition between document and equipment is an opportunity to lose place, skip a step, or misread a tolerance.

Low-frequency procedures compound the problem. A procedure performed monthly or quarterly is not retained the way a daily task is. The technician may have performed it four times in the last year. Memory fills the gaps between document consultations. That is where errors originate.

The document is not the problem. The format is. Paper and tablet procedures require attention to exist in a different location than the work. That split attention is the mechanism of error.

Aberdeen Group and PTC research on AR-guided maintenance deployments documents a 25 to 35% reduction in task completion time and a 40 to 50% reduction in error rate. Those outcomes share a cause: the procedure is at the point of work, not beside it.

What AR field guidance looks like in practice

AR overlays the procedure directly onto the physical equipment. An arrow appears on the display indicating the valve to open. A label shows the torque specification for the fastener the technician is holding. A highlighted zone marks the area to inspect. The technician advances to the next step with a simple gesture — a nod, a tap on a wristband, or a voice command.

The physical environment remains the primary space. The technician is looking at the equipment throughout. Nothing requires them to look away. The procedure travels with their attention, not against it.

Step-by-step guidance is the core capability. Every step in the procedure becomes a visual instruction anchored to the physical asset. Each step is completed in the field of view. The system tracks completion and prevents the technician from advancing past a critical checkpoint without confirmation.

NDA designs and deploys AR guidance systems for industrial field operations.

Use cases across field operations

Three use cases recur in industrial AR deployments.

Step-by-step procedure guidance covers the broadest ground. Any complex, multi-step procedure — PM routines, changeover sequences, calibration, inspection — can be delivered as an AR overlay. The value scales with procedure complexity and with how infrequently the task is performed.

Remote expert support handles the exceptions. When a technician encounters something unexpected — an anomaly in the equipment, an undocumented condition, a component that does not match the schematic — a remote expert session can be activated from the field. The expert sees exactly what the technician sees. The expert annotates in the technician's real space: a circle drawn around the anomaly appears overlaid on the physical component. The technician does not leave the job. The expert does not travel to the site.

Asset identification speeds every maintenance interaction. A technician points a tablet or smart glasses at a machine component. The system identifies the asset and pulls its maintenance history, open work orders, parts specifications, and relevant documentation from the CMMS — no manual lookup, no paper trail. The right information surfaces in the field, not back at the workstation.

Content authoring for maintenance AR

Deploying AR guidance requires content before hardware. That sequence matters.

Procedure content must exist in digital, structured form before it can be authored into AR. If the facility's procedures are in PDF binders or legacy document management systems, the first step is digitization and structuring — not AR authoring. This is typically the longest phase of an AR deployment.

Each step in a procedure must be mapped to a physical location on the asset: which component does this step address, what does the technician need to see, what spatial anchor ties the overlay to the physical object. Spatial anchoring requires reference imagery or 3D data for each targeted asset.

Content authoring tools — PTC Vuforia, Scope AR, or comparable platforms — allow subject matter experts to build procedures without software development skills. The output is a structured, versioned procedure that can be updated when the underlying process changes. Version control in AR content is not optional. An outdated step in an AR procedure is as dangerous as one in a paper manual.

Hardware selection for industrial conditions

Hardware selection is driven by the environment, not the preference.

Harsh industrial environments — high vibration, dust, oil, temperature variation — require purpose-built hardware. The RealWear Navigator 520 is designed for these conditions. It is head-mounted, hands-free, and rated for industrial use. It does not require a clean environment or careful handling.

Moderate industrial environments support tablet deployment. Tablets are lower cost, easier to manage, and carry familiar interfaces for technicians who already use them in their workflow. Tablets require a free hand, which limits use in tasks where both hands are occupied throughout.

Smart glasses — Microsoft HoloLens, Magic Leap, or equivalents — provide spatial anchoring and hands-free operation in controlled environments. They carry higher cost and greater fragility than purpose-built industrial hardware. Evaluate the environment first.

Battery life, connectivity requirements, and device management scale are the operational variables that determine feasibility, not device capability alone.

Measuring outcomes

Three measures matter: error rate reduction, time-on-task, and training cost avoidance.

Error rate reduction is measured against a documented pre-deployment baseline. Pull maintenance quality records from the CMMS. Identify error categories — steps skipped, torques outside spec, inspection items missed. Measure the same categories post-deployment. The comparison is the outcome.

Time-on-task reduction reflects the efficiency gain from eliminating document-to-equipment transitions. Measure before and after. High-complexity procedures show the largest gains.

Training cost avoidance is a function of how much technician time AR guidance displaces. If a new technician who previously required 20 hours of supervised practice before performing a procedure independently now requires 8, the delta is quantifiable. Apply the fully-loaded labor cost. That is a financeable number.

Integration with CMMS and ERP systems

AR guidance systems generate data: which technician, which asset, which procedure, what timestamp, which steps were confirmed or flagged. That data belongs in the CMMS — not in a separate reporting silo.

Integration with IBM Maximo, SAP PM, Infor EAM, or equivalent CMMS platforms allows AR session data to auto-populate work order completion records. The technician completes the AR-guided procedure in the field. The work order closes in the CMMS without a separate data entry step.

ERP integration extends this to parts consumption, labor reporting, and asset history. The field work and the record are synchronized. Plan the integration architecture before deployment. Retrofitting data pipelines after go-live adds cost and delay.

Mixed reality extends AR capabilities into remote expert support and spatial collaboration.