Spatial computing has had several false dawns. The difference now is that the hardware is capable, the tooling is mature, and enough enterprises have run real deployments that the pattern of where it pays off is clear. Meta Quest headsets are affordable and manageable at fleet scale, Apple Vision Pro has set a new bar for passthrough and display quality, and Android XR devices have joined the market. Meanwhile WebXR lets many experiences run in a browser with no app store involved.
The question for most organisations is no longer whether AR and VR work, but where they beat the alternatives on cost, safety or speed. At RixlSoft we build spatial applications in Unity, Unreal and WebXR, and this article sets out the use cases we see paying back, how to choose a platform, and how to structure a pilot that produces a clear go or no-go decision.
Immersive training: the clearest business case
Training is where virtual reality most reliably earns its keep, especially when real-world practice is dangerous, expensive, or hard to schedule. Think equipment operation, emergency procedures, hazardous environments, medical and clinical skills, and soft-skills scenarios such as difficult customer conversations. VR lets trainees repeat a procedure as many times as they need, fail safely, and receive consistent assessment.
The economic logic is straightforward. Compare the cost of the current method, including travel, instructor time, equipment downtime, consumables and incidents, with the cost of building and deploying the simulation. Programmes with many trainees, recurring certification or expensive physical setups tend to justify the investment fastest. Instrument everything: completion times, error rates and assessment scores give you evidence rather than anecdotes.
- Safety-critical procedures such as lockout-tagout, confined spaces and emergency response.
- Operating expensive or scarce equipment without taking it out of production.
- Onboarding for distributed workforces where central classroom training is costly.
- Scenario-based soft-skills practice with repeatable, scored outcomes.
Remote assistance and guided work
Remote assist puts an expert virtually beside a field technician. The technician wears a headset or uses a phone or tablet, the expert sees their view, and annotations are anchored in the real world. The value comes from fewer repeat visits, less expert travel and faster resolution. Step-by-step AR work instructions overlaid on equipment extend the same idea to routine tasks and reduce dependency on paper manuals.
Practical constraints matter more here than in training. Consider connectivity on site, safety rules about headsets in industrial areas, battery life across a shift, and whether a phone-based AR solution is good enough. Often it is, and it avoids buying and managing headsets entirely.
Product visualization and sales
For manufacturers, retailers, real estate and architecture, spatial visualization helps buyers understand products that are large, configurable or not yet built. Placing a true-scale machine on a factory floor, walking through an unbuilt apartment, or configuring a product in 3D shortens sales cycles and reduces misunderstandings that lead to returns or change orders.
The hidden cost is 3D content. CAD models are usually far too heavy for real-time rendering and must be optimised, textured and kept in sync with the product catalogue. Budget for a content pipeline, not just an app. Where reach matters more than fidelity, browser-based 3D and AR on mobile devices can deliver much of the value with no installs.
Choosing a platform: Quest, Vision Pro, Android XR or WebXR
Meta Quest devices remain the pragmatic default for training at scale: they are relatively inexpensive, standalone, support hand and controller tracking, and can be managed centrally with enterprise device management. Mixed reality passthrough on current models also enables blended scenarios. Check Meta's current business and device management options at the time of purchase, as they have changed more than once.
Apple Vision Pro, running visionOS, offers exceptional display and passthrough quality and integrates with the Apple ecosystem. It suits high-value visualization, design review and executive or client-facing experiences more than large training fleets, given its price. Apple also provides enterprise APIs for approved business use cases. Android XR, Google's platform with Samsung's Galaxy XR as the first headset, is a newer option worth watching for organisations standardised on Android and Google services.
WebXR deserves serious consideration. Experiences built with frameworks such as Three.js run in headset browsers and on phones, which removes app distribution and update friction. For lighter experiences, product viewers and pilots, it is often the fastest path. For demanding simulations, native builds in Unity or Unreal still win on performance and device features.
How to run a pilot that proves or disproves ROI
Most spatial pilots fail not because the technology disappoints, but because nobody defined success. Before building anything, pick one use case, one user group and one metric that matters to the business, such as time to competency, first-time fix rate or quote-to-order conversion. Capture the baseline from the current process.
Keep the first build small: one scenario or workflow, polished enough that users take it seriously. Plan the operational side as carefully as the software, including device charging, hygiene, storage, management, network access and support. Run the pilot for long enough to get past novelty, typically eight to twelve weeks, and compare results against the baseline and a control group where possible.
- Define one use case, one audience and one measurable outcome.
- Measure the baseline before building anything.
- Budget for 3D content and device operations, not just development.
- Collect comfort and usability feedback alongside performance data.
- Decide in advance what result justifies scaling.
Scaling beyond the pilot
Successful pilots scale when they are integrated into existing systems. Training results should flow into the learning management system, remote assist sessions should link to work orders, and product visualizations should pull from the same catalogue as the e-commerce site. Build a reusable content pipeline and component library so the second and third scenarios cost significantly less than the first. That is where the ROI compounds.
Plan for the human side of scale as well. Some users experience discomfort in VR, so keep sessions short, offer seated options and always provide a non-immersive alternative. Involve IT security early on device management, network access and data captured by cameras and sensors, especially for passthrough and remote assist, where video of facilities or people may be recorded. Finally, assign a product owner for the spatial programme. Without one, pilots tend to remain isolated demos rather than becoming part of how the organisation trains, sells and supports customers.
Key takeaways
- Training, remote assistance and product visualization have the most repeatable returns.
- Choose the device for the use case: Quest for scale, Vision Pro for fidelity, WebXR for reach.
- Define one metric and a baseline before building the pilot.
- Budget for 3D content and device operations, and integrate with existing systems to scale.
