Spark, Campfire’s AI assistant for physical work, shows guidance for removing a robotic dog’s top cover on a laptop, in a mixed reality headset view, and on a phone paired with AI glasses.

Insight

From Immersive Training to AI-Guided Work: How AI is Changing Enterprise XR

From Immersive Training to AI-Guided Work: How AI is Changing Enterprise XR

Nov 20, 2026

For years, enterprise XR teams have worked toward a familiar goal: make it easier for people to understand complex equipment and perform work with their hands.

Immersive training can help someone learn a procedure before working on the real equipment. But that same person may need guidance again next week, at a workstation, or months later while servicing a product. Their need for assistance continues, even as the setting changes.

An AI assistant gives companies a way to support those situations through one body of product and procedure knowledge. Immersive devices then become one way to deliver that assistance, as part of a broader AI solution for physical work.

Expert guidance beyond the expert-led session

An expert explaining a procedure does much more than read the instructions aloud. They point out the relevant parts, explain what’s happening, and answer questions as they come up.

Teams around the world already use Campfire for instructor-led training, with experts and learners exploring shared 3D models even when they’re in different locations. Of course, that still requires the expert to be there. If every walkthrough needs someone to lead it, their availability sets a limit on when people can receive that help.

Spark, Campfire’s AI assistant for physical work, turns a technical procedure into a conversation. It delivers approved procedures with verbal and visual guidance, while drawing on company knowledge to answer questions. So if a worker is removing a battery, they can point at a component on the model, ask what it is, and get answers about it from the product manual.

The expert’s role in defining and approving that procedure remains essential. What changes is that each worker can get responsive help while they follow it, without needing the expert to join them in real time. That makes the guidance useful well beyond a scheduled session.

Useful guidance doesn’t always need an immersive display

A headset makes sense when scale and spatial understanding matter. Being able to move around a life-size assembly and lean in to inspect its internals can make a complicated explanation much easier to follow.

Yet a technician performing a repair may need both hands for the work. A visual reference or video might still be needed as they complete a procedure, but the display doesn’t necessarily have to be inside a headset or glasses to be useful.

With AI glasses connected to a phone, the technician can ask Spark to explain the next step, repeat a detail, or bring up a relevant video. The glasses provide a hands-free way to have a conversation with Spark while they work, with the phone sitting nearby for visuals.

The technician can follow the procedure, navigate to the next step, and ask questions without picking up the phone. When they need a visual reference, they can glance at the screen. It’s a practical alternative to putting the entire experience into AR glasses when a nearby display is sufficient.

Companies can keep using immersive headsets where they add value without requiring every worker to use one to get assistance. The device can suit the task and environment, while the assistant provides continuity.

From manual authoring to expert review

Companies already have documented procedures for much of the work their teams do. The challenge is turning those existing instructions into interactive guidance an assistant can deliver, without a lengthy manual authoring process for every procedure.

A job sheet, for example, may already lay out each step of a procedure. But someone still needs to manually connect those steps to the 3D model, identify which parts to highlight, and decide what needs to move or be shown.

Multiply that preparation across many products and procedures, and the effort involved puts a practical ceiling on what a team can deliver.

AI can take on much of that work. Given an existing job sheet and 3D model, Spark can extract the documented steps and create visual guidance for each one. If an instruction says to remove the top-cover screws, for example, it can identify those components and highlight them on the model.

An expert still needs to preview the result, correct it, and approve it before anyone uses it. But their work shifts from manually building the entire guided experience to validating and refining what AI has already prepared. So the authoring effort becomes less of a constraint on which procedures a team can support, and how often they get updated.

A broader scope changes the business case

Let's say an XR team is focused on immersive training. Elsewhere in the same company, an AI team may be exploring how to make technical knowledge more useful to employees. An assistant for physical work connects those two efforts.

The business case now includes how the company uses AI to help people follow procedures, get answers, and work safely and productively in everyday operations. That can bring different objectives, sponsors, and budgets into a project that might previously have been approached as an XR initiative.

XR teams know how to explain complex equipment, connect instructions to specific components, and make guidance practical in a working environment. That experience matters when an AI team is deciding how an assistant should support physical work.

Getting involved early lets XR leaders help shape the larger solution, including where immersive technology adds value.

Explore Spark

Through the Spark Preview, Campfire is working directly with customers to turn their existing procedures into guidance Spark can deliver.

This article is adapted from a webinar hosted by the AR for Enterprise Alliance (AREA), featuring Campfire’s Jay Wright and AREA’s Mark Sage. The full recording is below.

Note: some details of the pre-release interface have been obscured in the recording.