Choosing the right technology
Virtual reality, augmented reality, mixed reality, AI and a fast build are not five points on one scale. They solve different problems, so the choice starts with the problem, not the hardware.
Start with the outcome
Consider VR when the goal needs full spatial immersion. Consider AR when digital information needs to sit over the real world.
Consider MR when physical and digital objects need to interact in the same space at once. Consider AI on its own when the core challenge is knowledge, language or conversation rather than 3D space.
Consider a fast build when you need to test an idea before committing to any of the above.
Each option has real limits, set out section by section below, and a number of real projects combine more than one.
Most of the options rule themselves out
Name the problem precisely and the shortlist gets short quickly. Naming the hardware first is what keeps it long.
The problem decides which door opens
Most enquiries arrive naming a headset, a chatbot or an app.
The result might need a room to be inside, guidance laid over the real thing, or digital objects the room can push back on.
It might need an answer in words, or a working version people can react to this fortnight.
Answer that first and the row below reads as a shortlist instead of a menu.
What does the outcome most need?
Questions about the result you want, and the technology each answer points to. Most projects start with one of these and add a second later.
When does virtual reality make sense?
Virtual reality replaces the user's physical surroundings with a complete 3D environment, seen through a dedicated headset.
When the room itself is the point
VR is worth considering when the physical environment is hazardous or costly, when the task needs complete visual and spatial focus, and when true physical scale matters.
It is less suitable when users need to handle physical paperwork, real tools or colleagues in the room, or when distributing headsets is impractical.
When VR fits
A place that is dangerous or expensive to practise in, a task that needs the whole of someone's attention, and a structure that only makes sense at its real size.
Where PTR's own evidence sits
Two healthcare programs where the conversation itself was the thing that could not be rehearsed safely any other way.
"As organizations tighten budgets and adapt to remote workforces, VR offers scalability, efficiency, and deep behavioural impact. Beyond training technical skills, it enables leadership, resilience, de-escalation techniques, and inclusive decision-making, all essential for thriving in today's workplace."
Simon Lowe
Co-founder, People Tech Revolution. Quoted in Forbes, June 2025
When does augmented reality make sense?
Augmented reality overlays digital information, such as text, diagrams or 3D models, onto a live view of the physical world through a phone, a tablet or lightweight smart glasses.
Guidance on top of the thing itself
AR is worth considering when people need hands-on guidance in context, when broad access matters more than depth, and when the real location is part of the experience.
It is less suitable for complex multi-user simulations that need full immersion or two-handed spatial interaction.
When AR fits
Steps shown on the equipment they belong to, a build that runs on the phones people already carry, and a place that has to stay in view.
When does mixed reality make sense?
Mixed reality blends the physical and digital worlds, so interactive 3D objects anchor to real tables, walls and floors through a passthrough headset.
Digital objects in the real room
MR is worth considering when people must handle digital models and physical objects at the same time, when training happens in person and in a group, and when situational awareness is critical.
MR needs a modern passthrough headset, so the initial investment is higher than mobile AR.
When MR fits
A hologram on a real desk, colleagues who can still see each other, and a room you must keep watching while you work.
When does AI make sense on its own?
AI systems process data, identify patterns and produce language without any 3D interface. They run on the web, a laptop or a phone, and no headset is involved.
When the problem is words, not space
AI alone is worth considering when the core problem involves text, language or data, when staff need to query complex guidelines at scale, and when the solution must run on the browsers and desktops an organisation already has.
PTR's own AI work centres on digital twins and conversational agents: versions of a real professional that answer questions from an organisation's approved knowledge, with guardrails, escalation paths and human oversight. AI overview
When AI fits
A problem made of documents and conversations, knowledge that people need to ask for rather than search, and a fleet of ordinary devices.
When does combining AI and XR make sense?
Put the two together and a character in a headset can respond to what a learner says, and a scene can change while it is being played.
A character that answers back
The combination is worth considering when learners need to practise interpersonal communication, and when a scenario has to adapt to what a person does in it.
It also adds development complexity, and it needs careful handling of voice latency and guardrails. PTR's VR programs pair branching scenarios with a live, human-facilitated debrief, the format behind every case study on this page.
Real-time conversational AI inside the headset exists today, and it is a newer, separate approach.
When AI inside XR fits
A difficult conversation that has to go somewhere unscripted, and a scene that should get harder or easier depending on how it is going.
When does a fast build make sense?
A fast build is an early working version of a concept, made to test specific assumptions, check demand or prove feasibility before anyone commits to a full production build.
A question, answered with a working thing
A fast build is worth considering when decision-makers need to experience a concept before committing budget, when there is real technical or design uncertainty, and when the timeline is tight: an event, a board presentation, a grant application.
A prototype exists to answer a question. The hardening a real rollout needs comes after it. Fast Builds overview
The five-step cycle
Brief, Build, Test, Refine, Ship. A prototype from two business days and a full build inside eight working weeks, so clients see working versions in sprints rather than one reveal at the end.
Comparison at a glance
The same questions asked of each option. Every column heading opens the page that explains that technology in full.
| VR | AR | MR | AI | Fast Build | |
|---|---|---|---|---|---|
| Real world visible? | No | Yes | Yes, and digital objects react to it | No headset | Depends on the prototype |
| Typical hardware | VR headset | Phone, tablet or glasses | Passthrough headset | Browser or desktop | Often existing devices |
| Several people in one room, seeing each other? | Not natively | Yes | Yes | Yes, it is not spatial | Depends on the concept |
| PTR example on this page | Mater, Mayo Clinic | Capability and demos | Capability and demos | Digital twins and agents | The Fast Builds process |
| Where it tends not to fit | Real tools, paperwork or colleagues needed | Complex multi-user immersion | Higher upfront cost than mobile AR | Physical skill or spatial reasoning goals | Final, fully scaled deployment |
What a choice looks like once it is built
Whichever technology you land on, the moment that matters inside a scenario looks like this: a person, a counter, and three ways to respond. Pick one, watch what it does, and go again.
Sector context
Each sector weighs the same options differently. Two documented PTR projects show how: broad, repeatable access in the public sector, and curriculum built into the session itself in education.
Questions to ask before you start
Work through these with your team before anyone names a technology. The answers usually make the choice for you.
The questions to ask
- What is the exact outcome you want? Physical skill, spatial awareness, interpersonal communication or data analysis?
- Who is the end user and where will they use it? At a desk, in a dedicated facility, in the field or at home?
- What hardware is already available? Existing laptops and phones, or a budget for headsets?
- What are your data privacy and security requirements? Does it handle sensitive personal data, and where must that data be hosted?
- Would a prototype reduce your initial risk? Can a focused build test the core assumptions before the complete system?
Hardware, deployment and governance
The practical questions that follow the choice: what you have to buy, who has to be in the room, and what rules the system runs under.
Where each technology is not the right fit
Every option above has a real limit. Knowing it in advance is cheaper than finding it in a pilot.
The short version
One rule covers the whole page: name the problem first, and the technology follows from it.
Start from the problem, not the technology
Start from the problem, not the technology. Full immersion points to VR. An overlay on the real world points to AR. Physical objects interacting with digital ones points to MR.
A knowledge or language problem points to AI on its own. An untested idea points to a fast build first.
Mater, Mayo Clinic, Central Coast Council, Etihad and Metro Trains are VR projects run with a human-facilitated debrief. PTR's measured evidence sits in VR.
The AR and MR work described here is capability and demonstration, and real-time conversational AI inside XR is a newer approach.
Talk to us about which technology fits your organisation.
Still deciding what fits your team?
Sources and further reading
Mater, Mayo Clinic and Google training evaluations, with guidance on AI risk management and government use.
Mater Immersive Learning Pilot closure report
The program evaluation behind the 93% and 96% figures quoted above.
Promoting Nursing Diversity, Equity, and Inclusion Through Virtual Reality Learning
The educational evaluation of the Mayo Clinic program described above.
A virtual reality intervention to increase interracial empathy and upstander behaviors in nursing leaders
The peer-reviewed, controlled study of a program offered to 1,149 eligible nurse and social worker leaders, with the 6 to 8 month follow-up.
Leveraging Virtual Reality to Enhance Diversity and Inclusion Training at Google
The published account of VR inclusion scenarios in a large workplace.
NIST AI Risk Management Framework (AI RMF 1.0)
The voluntary framework, built around govern, map, measure and manage, referred to in the governance card above.
Policy for the responsible use of AI in government, Version 2.0
The mandatory requirements and impact assessment tool for Commonwealth agencies.
