PsychProof
PsychProof Logo
CHANGE PLANNER

CHANGE MAY CREATE PSYCHOSOCIAL RISK. THE RECORD SHOWS WHETHER YOU MANAGED IT.

A restructure. An AI rollout. A new rostering system. Every code in Australia expects the psychosocial risk to be assessed while the change is still being planned. Change Planner does the assessment, shows you the gaps against the Code, ISO 45003 and the new digital work system duties, and seals the record.

17,600 claims

Serious mental health claims rose 14.7% in a year and 161% in a decade. Median cost $67,400, median time lost 35.7 weeks. Safe Work Australia, 2025.

23 July 2026

Safe Work Australia's AI and digital technologies guidance: work intensification, surveillance and unfair decisions by systems are WHS hazards, managed the ordinary way.

NSW, 2026

The Digital Work Systems Act adds a duty that work allocated by an algorithm, AI or platform must not put health and safety at risk. Entry permit holders will be able to inspect the system.

2 October 2026

People at Work closes. Its own review found "limited post-survey guidance for targeted action planning". The next step is the part that was missing.

HOW PSYCHPROOF HANDLES CHANGE MANAGEMENT

BUILT FROM THE EVIDENCE,
NOT A FRAMEWORK.

Plan a Change is a prompted entry into the same five-step pathway every hazard in PsychProof goes through: identify, consult, design controls, implement, review. It is built from longitudinal cohorts, systematic reviews and meta-analyses, not from a practitioner change-management model. Widely used frameworks like Kotter's eight steps or Prosci's ADKAR have no independent peer-reviewed evidence behind them, so neither shaped this design.

Change hurts, measurably, and the effect lasts longer than a 90-day review

  • A German employment survey of 53,295 employees, tracked from 2006 to 2018, found those who had been through change in the last two years were 1.82 times as likely to report poor mental health. The gap did not shrink across the twelve years of the study.
  • A Norwegian cohort covering five change types found employees who went through repeated changes in a two-year window fared substantially worse on every psychosocial measure than a single change did.
  • A Finnish cohort of 22,430 municipal employees followed for 7.5 years found major downsizing predicted higher sickness absence and roughly double the cardiovascular mortality among the people who kept their jobs.

A single review at 90 days measures the acute phase and misses the tail. Change Planner now asks how much change a cohort has already been through, and schedules follow-up reviews past the first 90 days for material changes.

Communication and participation solve different problems

A field study of workers facing organisational change found three kinds of uncertainty: strategic, structural and job-related. Communication reduced the first. Only participation in decision-making reduced the other two. Change Planner no longer treats them as interchangeable.

A realistic preview works

In a longitudinal field experiment after a merger, the plant given a realistic preview (what changes, what does not, the timeline, the criteria, the support) returned to pre-announcement trust. The plant that got the standard announcement never did.

Procedural fairness outweighs the outcome itself

A meta-analysis of 37 samples found procedural justice predicted how survivors of a downsizing responded more strongly than what they actually received. Change Planner records the decision rule and the account given to workers, not just the outcome.

The rollout itself is where most support is needed

A study of technology go-lives found units with more engaged peer helpers gained significantly more proficiency in the first six weeks. Adoption data from a major AI rollout showed a predictable dip in uptake between roughly week three and week ten. Change Planner's during-phase controls now cover both.

NEW IN THIS RELEASE

AI rollouts and digital systems now get their own check

AI or a new system? Four questions, not a new process.

If the change introduces or alters a system that allocates, monitors, scores, or decides about work, we ask what it does and whether it can create unreasonable workload, excessive metrics, surveillance, or unfair decisions - the record the NSW digital work system duty and the AI Safety Standard both ask for.

See the gaps before an inspector does.

Your plan is checked against the Code of Practice, your state's regulation, ISO 45003, and, for AI and digital systems, the guidance and duties above. Every row cites its source and shows covered, weak, out of order, or missing.

How much change has this team already absorbed?

Repeated change compounds. A cohort that's been through two or more significant changes in the last 24 months now raises the confidence of the demand and change-management hazards flagged, instead of every change being scored as if it were the first.

The preview and the decision rule, on the record.

For a redundancy, restructure, merger or AI rollout, the plan now checks for a realistic preview (what changes, what doesn't, the timeline, the criteria, the support) and, where it applies, the decision rule and the account given to workers - not just that a communication happened.

See the core steps

Architecture, cohort, hazards, controls, mitigation plan, evidence workspace — plus the digital work system profile and coverage check above for AI and digital changes.

Change Architecture

Change Architecture

Cohort Definition

Cohort Definition

Hazard Mapping

Hazard Mapping

Control Library

Control Library

Mitigation Plan

Mitigation Plan

Evidence Workspace

Evidence Workspace

A GAP, SHOWN

TWO OF ELEVEN. THAT IS WHAT AN AI ROLLOUT USUALLY LOOKS LIKE ON DAY ONE.

A 420-person organisation, 180 contact centre agents, an AI assistant that allocates tickets and scores calls. IT calls it well managed. Here is what the coverage report actually shows.

RequirementSourceStatus
Consult affected workers before the decisionWHS Act ss.47–49Out of order
Work allocation must not create unreasonable workloadNSW Digital Work Systems Act 2026; SWA AI guidanceWeak
Monitoring assessed for intrusive surveillanceSWA AI guidance, 23 Jul 2026Missing
Meaningful human oversight of the systemAI Safety Standard, Guardrail 5Missing
Workers can contest an AI-derived scoreAI Safety Standard, Guardrail 7Missing
At least one structural control (not training alone)SWA Code of PracticeMissing
…and five more requirements, covering records of the decision, the AI inventory, and the structural-control gate.

Four fixes close most of it: a human override on allocation, a human review and dispute route on scores, a named system owner with a review date, and the decision record attached.

Ten of eleven, with the consultation timing recorded honestly as a breach with mitigation. That one page is what goes to the board.

Fictional organisation. Real requirements — sourced below.

WE KNOW YOU'RE ALREADY DOING A LOT.

What you already have

You already have HR platforms, engagement surveys, and incident management systems.

  • Engagement platforms tell you how people feel right now.
  • Incident systems capture what went wrong after the fact.

What's missing

A documented process. Hazards identified before the change. Action plans with owners and deadlines, and a record you can hand to the board or an inspector. Timestamped before anything went wrong.

This is the compliance gap.

HOW THE AI EARNS TRUST

THE AI SHOWS ITS WORK.

Unlike a generic chatbot, every hazard and control PsychProof identifies cites where it came from — the Safe Work Australia Code of Practice, ISO 45003:2021, or the specific section of the WHS Act. Nothing is asserted without a source you can check yourself.

17

Hazards mapped

81

Cited controls

3

Evidence tiers

38

Requirements checked

HC-CHGHigh confidence

Poor Change Management

A significant change to work practices, roles and responsibilities without a managed process creates foreseeable psychosocial risk.

Evidence: WHS Act s.47 · SWA CoP 2024

Every hazard card in your report looks like this — traceable, not asserted.

SOURCES

WHAT THE LAW NOW SAYS.

Jul 2022, rev. 2024

Safe Work Australia model Code of Practice: Managing psychosocial hazards at work

Psychosocial hazards must be considered at the design stage of a restructure, and workers consulted before decisions.

2021

ISO 45003:2021, clause 8.1.3

Organisations need a process for communicating and controlling changes that affect psychological health.

2018

ISO 45001:2018, clause 8.1.3

Planned changes are assessed for OH&S impact before implementation.

1 Dec 2025

Victoria, OHS (Psychological Health) Regulations 2025

Training or information cannot be the main control. A change to work triggers a review.

18 Feb 2026

NSW, WHS Amendment (Digital Work Systems) Act 2026

Work allocated by algorithm, AI, automation or platform must not put health and safety at risk. Commencement by proclamation.

1 Jul 2026

NSW, WHS Act 2011, s.26A

Compliance with the Code is a standalone enforceable duty.

23 Jul 2026

Safe Work Australia, AI and digital technologies guidance

Intensification, surveillance, fatigue and discriminatory decisions from digital systems are WHS hazards.

Sep 2024

Voluntary AI Safety Standard, ten guardrails

Risk management, human oversight, contestability, an AI inventory with owner and review date, stakeholder engagement.

2 Oct 2026

Safe Work Australia, People at Work decommissioning

The survey closes; organisations need a next step for the results.

2025

Safe Work Australia, Key WHS Statistics 2025

17,600 serious mental health claims, up 14.7% year on year and 161% on the decade.

WE CAN'T PROVE YOUR PAST DECISIONS. WE CAN SEAL THE FUTURE ONES.

Can't our WHS consultant just do this?

They can, and for a major change you should have one. Change Planner does the mapping and the coverage check in an afternoon, so your consultant's time goes on the interviews, the root causes and the implementation, not the paperwork. PsychProof works with a network of independent WHS consultants who use the platform with their clients.

Is this an ISO 45003 certification?

No. It measures your change plan against the clauses that apply to change and shows the gaps with sources. Certification is an external audit. This gives you the evidence an auditor asks for.

We're in NSW and rolling out AI. What changes?

The Digital Work Systems Act adds a duty about work allocated by AI, algorithms and platforms. Change Planner asks the four questions the Act captures, records the system owner and oversight point, and shows the rows as pending commencement until the Act is proclaimed.

Does this replace our HR system?

No. This sits alongside it. HR systems manage the employee lifecycle. PsychProof Change Planner creates the forensic evidence that you proactively managed WHS risk during operational changes — something no HR platform is built to do.

How accurate is the AI — could it get something wrong?

The coverage table isn't AI. It's a set of rules over what you recorded. Only the recommendation text is written by AI, and every hazard, control and requirement cites its source (the Code of Practice, ISO 45003:2021, or a WHS Act section) with an evidence tier, so you can see exactly how strong the basis is. It's a structured starting point for your review and consultation, not a replacement for your judgement or your workers' input.

What does it cost?

Change Planner is part of the PsychProof platform. See current plans and what's included at each tier on our pricing page.

Is our data kept private?

All customer data is stored and processed exclusively in Australia (Sydney) — never offshore. Full detail on encryption, access controls, and our security posture is on our Trust page.

We've already had an incident related to this kind of change — does this still help?

Yes. A documented assessment doesn't undo what's already happened, but it's the difference between having a record of genuine effort going forward and having nothing to show a regulator next time. Most organisations that adopt Change Planner do so after a near-miss, not before one.

Get a Sample Report

Rolling out AI or restructuring this quarter?

Send us the outline and we'll return the coverage table for it — the gaps against the Code, ISO 45003 and the digital work system duties, sourced.

Try Our Pilot Programme