AR instructions for industrial equipment

We start with an online meeting with maintenance and operations staff: chief engineer, area supervisor, service team. We work out: which equipment the instruction is for, which operations we cover (installation, maintenance, repair, diagnostics), who the target user is (own staff, contractors, new hires) and what should change after rollout.

The client sends us source materials in electronic form: equipment drawings and technical documentation, photos from different angles, description of markers and working environment, operation regulations. Output — a 5-7 page brief with scenario descriptions and reporting requirements.

We map out each operation step-by-step: what to do, with what, in what order, how to verify the result. We align the sequence with your technical staff and safety team — what's critical, what can be skipped, where the risks are.

We prepare a short pilot instruction (one or two operations, 5-10 steps) to validate the concept. It's easier to get feedback on this than to rework a finished product.

We model equipment from your drawings and technical documentation: precise dimensions, details, elements to attach hints to. We set up recognition — by marker (sticker, QR code), by object shape or by environment photogrammetry, using your photo materials.

Precise tracking is what tells a working AR instruction apart from «floating» hints. We test recognition stability under different lighting, viewing angles, metal glare — on the materials you provided and via remote tests in your environment.

We develop on Unity — a universal engine that works with most AR devices: iOS tablets (iPad, iPhone Pro with LiDAR), Android tablets and smartphones with ARCore, Microsoft HoloLens 2 AR glasses, Magic Leap 2, RealWear (for harsh environments).

One build — all devices. The instruction works offline (important in workshops and remote sites), syncs when connected. We integrate with your maintenance system, MES or ERP via API — operations are logged into shared reporting automatically.

We gather a remote pilot group of your employees on one area, send them the build and setup instructions. We help via online calls with device setup, train instructors and first users, run a cycle. We collect feedback: where it's inconvenient, unclear, where tracking «floats» — and fix immediately.

After the pilot we roll out to the whole enterprise and hold active support for the first 6-8 weeks: answer user questions, add missing steps, refine complex spots based on practice.