100%
Match with manual screening
Across every candidate in the pilot
ITAC 2026 Finalist · Data InsightsPredicting who will stay, before the offer is made.
Feros Care replaced a 60-minute psychometric test with a five-minute behavioural assessment, and it reached the same hiring conclusions as their most experienced recruiters, every single time.
10 million cups of tea. Countless moments shared between the same resident, same carer.
30–50min
Saved per candidate
Screening time returned to the team
~200
Candidates assessed
During the validation pilot
The challenge
Early departure was rarely about capability.
Every person who receives care deserves a worker who genuinely wants to be there: someone who knows their name, understands their routines, and has chosen this work because it matters to them. For older Australians and people living with disability, that continuity is foundational to dignity, safety and trust.
Feros Care's discovery phase revealed a consistent pattern. People did not leave because they couldn't do the job. They left because of an expectation-reality gap.
Workers arrived expecting warmth and purpose. They met administrative load, emotional exposure, and complexity they had never been screened for. The psychometric tool in use took up to 60 minutes to complete, long enough to deter the very candidates the sector needs, and it could be gamed by anyone who knew the socially acceptable answer.
Long-tenured Feros employees, by contrast, shared identifiable traits: values alignment, emotional intelligence under pressure, psychological flexibility, and a capacity to find meaning in human connection. Those traits are measurable. They simply weren't being measured.
What we built
Attitudinal science, not personality testing.
Personality tools are gameable. ethyx measures genuine motivational commitment and behavioural intent, backed by tamper-tracking and inauthentic-input detection, so it surfaces what a candidate actually thinks and is likely to do, not what they believe the recruiter wants to hear. The model was built from real Feros Care workforce data: interviews with people who stayed, people who left, and the managers who knew both.
Layer 01
Risk scoring
Attitudinal risk scoring and dropout prediction, placing each candidate on a commitment spectrum before the offer.
Layer 02
Sentiment analysis
AI sentiment analysis provides objective behavioural signals during screening, removing the variability of gut feel.
Layer 03
Interview intelligence
Dynamic, candidate-specific interview questions equip hiring managers to probe the dimensions that matter for each person.
And it learns. Once a candidate is hired, their real retention trajectory feeds back into the model, recalibrating its scoring on Feros Care's own outcomes. It is not a static test. It is a living system that grows more accurate with every cohort.
The results
What the pilot showed.
100%
Concordance with expert judgement
Across every candidate in the pilot, ethyx reached the same conclusions as Feros Care's experienced HR team, in a fraction of the time.
30–50min
Saved per candidate
Time previously spent on paper and telephone screening, handed back to the team for onboarding, culture and the people already in care.
A 10% reduction in turnover would save $320,000 to $450,000 a year across a 100-staff team.
Preventing a single early departure pays for the entire programme. Feros Care has already seen a meaningful reduction in turnover in the months since deployment. Financial case per SHRM and Deloitte benchmarks; current data available on request.
This is the figure behind the ten million cups of tea. Not money saved, but care made continuous for the people who depend on it.
Who benefited
The outcome that matters most isn't on a spreadsheet.
The HR team
A significant drop in time-per-hire and decisional burden. Interviewers now arrive with candidate-specific intelligence, shifting from information gatherer to relationship builder.
Candidates
A short, respectful, mobile-friendly assessment that treats applicants as people, not paperwork. Removing the 60-minute barrier keeps candidates the sector cannot afford to lose.
The people in care
When the right worker is hired and stays, a care recipient experiences consistency. They are known by name. Their preferences are remembered. Their trust is earned over time.
How it was built
A genuine co-design, not a vendor delivering a product.
This initiative was shaped at every stage by the people closest to the problem. Feros Care's recruitment team provided the ground truth, and their expert judgement became the benchmark, which is why concordance reached 100%.
Stacey Appleby
Executive Manager Disability & Community Development, Feros Care
Executive sponsor and strategic champion. A member of ethyx's Customer Advisory Board, ensuring Feros Care has an ongoing voice in the product roadmap.
Feros Care Recruitment Operations Team
Implementing partner
The source of the ground-truth data. Their expert manual screening judgements became the benchmark the model was validated against.
Dr Darren Coppin
CEO & Chief Behavioural Scientist, ethyx
Architect of the predictive model, drawing on a PhD in behavioural change and a decade of research across hundreds of thousands of participants in seven countries.
Devon Passmore
Director of Product, ethyx
Translated the behavioural science into a platform frontline HR teams actually use, on the principle that the best software feels invisible.
What people said
“What I love about this is it's using behavioural science and AI. It identifies candidates most likely to stay before they're even hired.”
Julie AndersonHead of Innovation, Ageing Australia
From a recruitment advisor who was, by their own admission, initially sceptical:
“I fear this tool is great.”
Feros Care recruitment advisor
The partnership
Invited to speak at Grow Bold.
The partnership worked so well that Feros Care invited us to present at their Grow Bold Conference, sharing the science behind ethyx with the people who care for older Australians every day.