TL;DR
Why now
- Legal mandate: directors carry personal criminal liability for psychosocial risk under modern WHS law
- Economic pressure: psych claims are the fastest-growing WC category — 161% growth in a decade, 38% of total cost
- Empty category: no engineering controls exist for this risk class
What Echo does
- Autonomous voice check-ins → structured signal → manager action → board-grade audit trail
- One closed loop. Running every week. For every team.
- Deploy in 14 days, board-ready inside 90
Traction
- Live in pilot with paying design partner
- ~50% voluntary enrollment, ~70% fortnightly coverage
- Workers consistently rate Echo 5/5 for trust
Why now
Three structural forces have converged to create a window that did not exist 24 months ago:
1. The positive duty
All Australian jurisdictions now require employers to identify, assess, and control psychosocial hazards with the same statutory rigour applied to physical safety. Directors face personal criminal liability — including imprisonment — for reckless failure.
This is not discretionary compliance. It is board-level governance, enforced with the full weight of WHS prosecution.
2. Rising claims, rising premiums
- 17,600 serious mental health claims (2023–24), +14.7% YoY
- +161% growth over the decade — fastest-growing category
- $67,400 median payout per claim (4.1× physical injury)
- 35.7 weeks median time lost (5× physical)
- 38% of total WC cost — from only 12% of claims
- $1.19 billion direct compensation annually
3. No engineering controls exist
- Surveys — lagging indicator. Measures last quarter's sentiment, not this week's risk.
- EAPs — reactive. Post-incident safety net. No detection capability.
- Consultants — point-in-time snapshots. Slow and expensive.
The market is buying fire insurance, but nobody is selling a smoke alarm.
The problem
Psychosocial risk is invisible, expensive, and structurally uncontrolled.
Frontline and shift workers — in mining, logistics, healthcare, construction, energy — operate under cognitive load that no walk-around can see. Workload pressure, roster conflict, personal strain, bullying, fatigue. These hazards don't appear in incident reports until someone is hurt, burned out, or gone.
Managers lack the signal. They rely on gut feel, toolbox talks, and whatever a worker is willing to say in front of the crew.
Boards are exposed. They carry a positive duty they cannot discharge with policies alone. Regulators are explicit: tick-box exercises, awareness training, and EAPs do not satisfy the obligation to actively control psychosocial hazards.
How Echo works
A closed-loop operating system that detects psychosocial risk, delivers actionable intelligence, closes the loop with managers, and generates board-grade evidence — autonomously, every week.
1. Check-in
Workers receive a scheduled voice call on their mobile. No app, no download, no login. Echo remembers previous conversations. Workers rate the experience 5/5 for trust and describe it as "talking to a real person."
2. Signal
Every call produces structured signal: themes, risk indicators, emotional context, operational pressures — extracted from natural voice. Echo doesn't need 100% participation. Like seismology, it triangulates from partial data.
3. Insight
Managers receive a weekly Vibe Report by email — roster-aware, action-oriented, specific to their team. Workers receive a private monthly Reflection Receipt. Operations leaders receive a Cohort Vibe Report — a Monday-morning command view.
4. Action
Tiered suggested actions: a 2-minute move, a 10-minute move, and a deeper move. Named people, specific times, practical steps. Every action is logged — receipt confirmed, taken, deferred, or dismissed.
5. Evidence
The audit trail connects detection to notification to action to outcome. Board-level evidence packs generated automatically. This is the core proof of "reasonably practicable" risk control for regulators.
AI-native, not AI-enabled
Purpose-built architecture: voice-first delivery, autonomous orchestration, Trait × State × Context modelling, and grounded in Psychosocial Control Theory (PCT) — co-developed with Prof. Warren Mansell, a global authority on PCT.
Privacy is architecturally enforced
Echo is a filter, not a microphone. Managers never hear raw audio. They see aggregated, anonymised cohort trends. Cohorts below 7 are suppressed. Workers control what they say, when they say it, and whether to share their receipt.
Traction
Echo is live in pilot with a paying design partner. Real workers, real calls, extended deployment.
Engagement metrics
- ~50% voluntary enrollment of eligible workforce
- ~70% fortnightly coverage of enrolled workers
- 6–8 min average call length
- 5/5 trustworthiness rating — consistent across all post-call surveys
- Zero negative feedback to date
Tangible outputs in production
- Weekly Vibe Reports — delivered to team and area managers
- Monthly Worker Reflection Receipts — delivered privately to workers
- Cohort Vibe Reports — cross-site intelligence for operations leaders
- Individual and Team Personality Reports — BFI-2–based, role-contextualised
- Board-ready audit trail artifacts — foundational data pipeline live
What workers say
"It almost felt like I was talking to a real person. He even took the initiative to schedule a follow-up call later this week."
"His clarity questions and curiosity had me reach a new level of insight."
"I LOVED my call — he's so sweet. It's super important to leave someone with a 'next step' rather than just being heard."
"The Weekly Vibe Report confirms my intuitions and makes me more confident in the actions I take with the team." — Area Manager
What this proves
- Workers will talk to an AI agent voluntarily, on shift, at sustained rates. This was the core product risk and it is de-risked.
- The signal is rich, specific, and actionable. Managers report it confirms their intuitions and directs their attention.
- Voice-first eliminates the friction barrier that kills survey participation in deskless workforces.
- Trust is the product. 5/5 ratings are a prerequisite for everything else. Without trust, the flywheel doesn't turn.
The frontline economy: massive, regulated, hiding in plain sight
🦺 Built for the 2.7 billion workers AI will not replace
"AI excels at writing and information work, but is less suited to tasks involving physical activity." — Microsoft Research (2025)
While white-collar tasks are being automated, the world's 2.7 billion deskless workers — on rigs, in mines, hospitals, farms, and logistics — are not going anywhere. These are the jobs that demand presence, judgement and coordination — not prompts.
"By 2030, up to 30 % of hours worked could be automated — meaning 70 % of human work remains deeply human." — McKinsey Global Institute (2023)
Developed markets at a glance
- Frontline & high-risk workers: ~140–180 million
- Annual payroll tied to these roles: ~US$5–6 trillion
- Employers with ≥500 frontline staff: ~40–60k
Scope: US, UK, EU, CA, AU, NZ. Roles include mining, energy, logistics, manufacturing, construction, healthcare, aged care, utilities.
Why this spend is durable
- Legal duty of care: psychosocial obligations are explicit and rising
- Workers' compensation premiums: direct cost reduction for the employer
- Insurance & audit pressure: boards must evidence live controls, not intent
- Operational stability: fatigue, attrition, and incidents have direct P&L impact
Echo's budget line is not "HR discretionary." It sources from legal, insurance, and operational stability — compulsory spend.
Adjacent market context
- Employee engagement software market: ~US$1.05B (2024), growing at 16.4% CAGR
- Echo's wedge — psychosocial risk compliance — accesses the larger EHS market (~US$7–9B)
- Then expands into engagement as a proven capability
Where the workers are (illustrative ranges)
Healthcare & Aged Care
- ~20–35m clinicians, nurses, carers
- Consumer-safety sensitivity; shift fatigue
Logistics, Transport & Distribution
- ~15–25m drivers, warehouse, last-mile
- Fatigue, distraction, linehaul isolation
Manufacturing & Fabrication
- ~30–50m operators, technicians
- High-tempo floors; error costs compound
Construction
- ~12–20m trades, site crews
- Permit pressure; subcontractor complexity
Energy, Utilities, Oil & Gas
- ~3–6m field crews, plant ops
- Critical lifts/permits; remote work
Mining
- ~1–3m pit & processing roles
- Rostered isolation; heavy mobile equipment
Notes: figures are rounded ranges compiled from national labour statistics and industry reports; they serve to show order-of-magnitude only. Detailed, sourced breakouts appear in the data room.
Why this compounds
Proprietary data flywheel
Every check-in, intervention, and outcome labels the model. More sites → richer labels → better interventions → measurable outcomes → higher value per seat. Early sites seed benchmarks that later sites rely on.
Regulatory embeddedness
Echo's audit trail — detect → notify → action → evidence — becomes the evidence of compliance for ISO 45003. Ripping out Echo means ripping out the audit trail. Structural switching costs anchored to the board's legal obligation.
Workflow gravity
Echo is not a dashboard people check. It is the system that calls workers, generates manager cheat sheets, logs actions, and produces board packs — autonomously. If Echo stops, the safety control loop stops.
Frontline engagement edge
Voice-first delivery, tone, turn-taking, and conversation cadence tuned for FIFO and shift workers in their vernacular, on devices they already use. Hard to replicate without the operational context.
Scientific foundation
Grounded in Psychosocial Control Theory (PsCT) — co-developed with Prof. Warren Mansell, a global authority on Perceptual Control Theory. A single, unified, testable mechanism for wellbeing, safety, and performance.
Why incumbents can't replicate
Survey tools (CultureAmp, Qualtrics) are architecturally different — forms, not voice. EAPs are reactive by design. Foundation model vendors lack the domain data and regulatory embeddedness.
Go-to-market
Beachhead: Western Australia
- Mature WHS regulatory environment with active enforcement
- Concentrated heavy industry (mining, energy, logistics) with large FIFO/shift workforces
- High psychosocial risk exposure and rising premiums
- Cultural familiarity with safety systems and compliance disciplines
Operating promise
- Deploy in 14 days on pilot sites
- Board-ready inside 90 days
- Zero IT integration required to start — roster upload + phone calls
GTM motion: Wedge → Land → Expand
- The Wedge (Months 1–6): Paid pilot cohort. Design-partner governance. Broker-supported introductions and safety leadership forums.
- The Land (Months 7–12): Proven pilots convert to full site rollouts. New paid pilots continue launching.
- The Expand (Months 13–24): Lateral expansion — existing customers add second and third sites. Channel partner deployment. Partner-led site deployments.
Global expansion: beyond Australia
Echo's regulatory wedge is not Australia-specific. Psychosocial risk legislation is converging globally. Three English-speaking markets have equivalent regulatory frameworks, director liability exposure, and frontline workforce concentrations — and require minimal product localization:
- 🇳🇿 New Zealand: HSWA 2015 director duties virtually identical to AU model. WorkSafe guidance released 2025. Natural adjacency — many customers will already have NZ sites.
- 🇨🇦 Canada: Quebec mandated psychosocial risk prevention (Bill 27, Oct 2025). Alberta covers PTSD for all workers. Oil sands culture mirrors WA mining — same buyer persona.
- 🇬🇧 UK: HSE integrated psychosocial risk into routine inspections 2025–26. 17M working days lost to stress per year. 8M construction workers, NHS, North Sea energy.
- 🇺🇸 USA: No federal mandate — but state-level psych claim expansion (NY 2025) and insurance/litigation economics (psych claims cost 2.5–6× physical) create equivalent buyer urgency.
The same failure modes — shift-work fatigue, rostered isolation, psychosocial risk from cognitive overload — present identically from Perth to Pittsburgh to Aberdeen. Australia is the hard case. Every other English-speaking market is additive.
Product roadmap
Board-readiness is the wedge. Subsequent modules compound the value of the data already being captured.
Board-Readiness Pack Now
ISO 45003 control coverage, hazard heatmap, intervention log, quarterly board pack and audit trail.
Attrition Risk
Risk flags, supervisor prompts, and replacement forecasting to stabilise teams.
Fatigue Index
Roster risk scoring, alerting and mitigations tuned to shift patterns and site context.
Scheduling
Team design and shift optimisation using human-factor signals and constraints.
Financial Impact Mapping
Jurisdiction-specific frequency/severity trends and premium trajectories as data matures.
Psychosocial Control Theory: the science behind Echo
Co-developed with Prof Warren Mansell, School of Population Health, Curtin University.
"Psychosocial risk is best understood as loss of control originating from goal conflict in workers."
Echo is built on Psychosocial Control Theory (PsCT), a unified model for workplace safety and well-being that explains how conflicting demands ("hit the quota" vs "do it safely") create measurable stress, fatigue, and unsafe adaptation.
- Deep roots, not pop-psych: PsCT sits on Perceptual Control Theory (Powers, 1973), one of the most rigorously modelled frameworks in behavioural science, validated through computer simulations, lab tracking tasks, and clinical use in Method of Levels therapy.
- From ISO 45003 lists to mechanisms: The white paper maps standard psychosocial hazards (workload, lack of control, bullying, isolation, traumatic events) to concrete control-loop failures leaders can measure and fix.
- Operationalised in Echo: Voice check-ins detect early signs of goal conflict and loss of control, quantify perceived control, and generate ISO-45003-aligned, auditable evidence of live psychosocial risk management.
Team
Two full-time co-founders. Senior advisors. AI-native operations — revenue scales with seats, not headcount.
Fletch Young · Founder / CPO
Ex-Uber: launched WA to nine-figure GTV. BHP Mining & Met ops → Strategy & BD. Enterprise healthcare AI strategy. Eng/Sci; LBS MBA.
Leonardo Fernandez Sanchez · CTO
Ex-McKinsey Digital: rebuilt online ordering platform for global pizza delivery chain — high-throughput, real-time, zero-downtime. Ex-Unilever Digital Ventures CTO. Deep cloud engineering and SRE discipline.
Mark Heath · Commercial Advisor
Ex-Goldman IBD → Uber ANZ leadership → Sequoia-backed founder; now COO at an ag-tech scaleup. Fundraising and enterprise BD.
Prof. Warren Mansell · Psychology Advisor
Global authority on Perceptual Control Theory; BA Cambridge, DClinPsy KCL, DPhil Oxford. Guides psych modelling and intervention logic.
Tom Rayner · Advisory Board Chair
Partner, Deloitte Risk Advisory (Perth → Sydney). Focus on governance systems in large public companies. Client services lead for Woodside.
Risks & mitigations
"Surveillance" pushback
Radical transparency, voluntary opt-in, anonymised cohorts, worker-benefit-first design. 5/5 trust ratings from live pilot validate the approach. Privacy is architecturally enforced, not policy-dependent.
Cold start without claims data
Start with compliance outcomes and early proxies (participation, manager action rate, signal coverage). Connect to frequency/severity as data matures. The audit trail has standalone legal value.
Feature replication by incumbents
Survey tools are architecturally different (forms, not voice). Echo's moat is in the data flywheel, regulatory embeddedness, and frontline engagement tuning — not features.
Model dependency / AI commodity risk
Echo uses multiple model providers. Value accrues in the orchestration layer, proprietary data, and domain-specific fine-tuning — not in any single foundation model.
Regulatory complexity across jurisdictions
WA beachhead establishes the governance, privacy, and compliance playbook. Psychosocial risk legislation is converging globally. The WA pilot is the hard case; other jurisdictions are additive.
AI ethics and responsible AI
PsCT provides a principled framework: if Echo undermines worker control, it replicates the harm. Design principles grounded in decades of clinical research, not ad hoc ethics guidelines.
Investor FAQ
Where does Echo start and expand?
Western Australia heavy industry first, then English-speaking markets with equivalent regulatory frameworks (NZ, Canada, UK). The same psychosocial risk failure modes — shift-work fatigue, rostered isolation, cognitive overload — present identically across developed markets.
How do you protect worker privacy?
Two-stream architecture: private coaching to workers; anonymised cohort analytics to management. Cohorts below 7 are suppressed. Tunable dials for retention, anonymity thresholds and safety exceptions. Privacy is architecturally enforced, not policy-dependent.
What happens if a Big Tech platform builds this?
Foundation model vendors don't have the domain data, the regulatory embeddedness, or the frontline engagement tuning. Survey tools (CultureAmp, Qualtrics) are architecturally designed around forms, not voice. Echo's value is in the closed-loop operating system — the signal-to-action-to-evidence chain — not any single AI capability.
What is the business model?
Per-seat subscription comparable to leading EHS suites. Fixed-scope pilot with 90-day exit. Modules layered post-pilot (attrition, fatigue, scheduling). Outcome-based components only after independent validation.
Is the product truly AI-native, or an AI wrapper?
Echo is purpose-built: voice-first delivery, autonomous call orchestration, Trait × State × Context modelling, and grounded in Psychosocial Control Theory. Core processes — call scheduling, signal extraction, report generation, action tracking, evidence logging — run autonomously. Revenue scales with seats, not headcount.
Let's talk
We're building the category-defining control for psychosocial risk in frontline workforces. If you invest in vertical AI, compliance infrastructure, or the future of work — we'd like to meet.
- Data room available
- Demo conversations live
- Board pack samples on request
For intros or materials, contact us.