Measuring how people think, inside the scene.
PeakXR is an R&D company building talent assessment for spatial computing. Participants step into realistic work scenes on Apple Vision Pro. Our models turn what they do there into scored, population-normed competency profiles.
Assessment still runs on paper and opinion.
Personality inventories, role plays scored by a person with a clipboard, and case studies on a desk all capture what people say they would do, filtered through whoever is watching. None of them scale well, and few hold up in hybrid and remote organisations.
| Parameter | Assessment centre | PeakXR |
|---|---|---|
| Input | Self-report, interviews, observer notes | Behaviour recorded inside the scene: choices, timing, movement, speech |
| Scoring | Assessor judgement on the day | The same models applied to every participant, normed against the population |
| Consistency | Depends on who runs the session | Identical scenario, identical hardware, identical scoring |
| Throughput | A handful of people per room, per day | Parallel sessions, one headset each, in any meeting room |
| Equipment | Dedicated room, trained assessors | Apple Vision Pro. No wearables, no controllers |
From behaviour to a score.
Every number in a PeakXR report can be traced back through the pipeline to something the participant actually did in a specific scene.
Scenario engine
A realistic work situation with characters that react to the participant: an office, a meeting, a field task.
Interaction events
Head and hand movement, what is looked at and touched, choices, timing and spoken responses.
Behavioural features
Raw events become features: hesitation, revisits, plan corrections, turn-taking, recovery after errors.
Scoring & norms
Machine-learning models map features to metrics, subdomains and domains, then place them against population norms.
Reports
An individual report with expert review for every participant, and a SARF view for the organisation.
Two modules. Domains, subdomains, metrics.
The measurement model is hierarchical. Metrics roll up into subdomains, and subdomains into domains, so results are comparable across people, roles and departments.
Cognitive
7 domains17 subdomains
47 metrics
How a person takes in information, holds it, and turns it into decisions under realistic pressure.
- Attention
- Visual and visuospatial attention efficiency
- Working memory
- Efficiency, visuospatial and verbal capacity
- Cognitive flexibility
- Hysteresis and blockade: holding on to a plan that has stopped working
- Problem solving
- Knowledge and source usage, solving efficiency
- Planning
- Goal-directed thinking, loop susceptibility
- Reasoning
- Analogical, visuospatial and causal reasoning
Socio-emotional
10 domains24 subdomains
80 metrics
How a person behaves with other people, especially when the conversation gets difficult.
- Communication
- Expressiveness, questioning, respectfulness, adaptability
- Empathy
- Cognitive and behavioural empathy
- Assertiveness
- Defending one's position, independence, directiveness
- Conflict management
- How harm to each party is judged, and how that shapes resolution
- Interpersonal cohesion
- Teamwork, fairness, unbiased approach
- Resilience
- Hardiness, positive future orientation, self-efficacy
A report a person can actually read.
Each participant gets an overall score for every module, domain and subdomain, placed on a population scale, together with a written expert review that explains what the pattern means at work.
HR gets the same data rolled up across teams and job groups in the SARF executive report.
Above population average
Headset on to report.
No prior headset experience is needed. There are no controllers to learn: participants look, reach and speak the way they would in the room itself.
Fit & calibrate
The headset is fitted and calibrated, then a short practice scene introduces the interactions.
Scenarios
Between 4 and 12 VR experiences, each built to draw out specific competencies.
Scoring
Behaviour is scored automatically against population norms. Nothing is marked by hand.
Report
Participants receive their report through a private, password-protected link. HR receives the team view.
A role play in a meeting room is still a role play. A tense client meeting you are standing in is a different test.Ecological validity — measuring people in settings that behave like the real world
Run a pilot with one team.
We bring the headsets, run the sessions on site, and walk you through the reports afterwards.