Trumis TRUST & PROMISE
Menu
Trumis / Advocates
Long-form research in the public interest

Advocates

Trumis sits at the point where a complaint is made, which is an unusually good place to see what happens to it afterwards — and what happens is not well documented anywhere a member of the public can read. Advocates is where we publish the long-form work.

We publish it because the people most affected by how complaint systems behave are the least able to see inside them. Advocacy, here, means putting the evidence where they can read it — not asserting a conclusion on their behalf.

103
Primary references across both reports
11
Evidence streams, adversarially tested
Per claim
Confidence marked High / Medium / Low
What you can expect from a report here
01 Primary sources, linked.

Legislation, regulator guidance, tribunal decisions, ombudsman annual reports and standards — cited so you can check them, not summarised so you have to trust us.

02 Confidence marked per claim.

High, medium or low, shown inline, so a well-evidenced finding never borrows authority from a thin one standing next to it.

03 Counter-evidence given its own section.

Every report includes what the record does not support, including where it cuts against our own thesis.

04 Corrections recorded, not deleted.

Widely-circulated figures that failed verification are listed with the reason, so the report can be used without re-litigating them.

Published

2 reports · 103 primary references · corrected in place and dated
Report 01 25 July 2026
Working research

The complaint that cannot land

What actually changed between 2010 and 2026 in how Australian organisations answer for the conduct of their frontline staff — and why the customer’s experience of being treated as the problem is real, even though almost nothing in the record says organisations decided to take the employee’s side.

The report tests a specific claim across six parallel evidence streams, one of them tasked solely with refuting it. The claim breaks in half.

How the claim breaks
Fails as stated
No authority tells an employer to back the worker.

No Australian authority tells an employer to back a worker on the merits of a customer’s complaint, and accountability for frontline conduct has increased since 2015 in every sector touched by a royal commission.

Survives, restated
The individual complaint was decoupled from the system.

Accountability migrated upward to regulators and named accountable persons, while the first-line channel an ordinary customer could reach was thinned by cost, offshoring, franchising and automation.

Mechanism
A ready-made doctrine supplied the vocabulary.

The Unreasonable Complainant Conduct framework, built from 2006 for a rare and genuinely difficult minority, became a licence condition for every financial services firm in 2021 and normative for everyone else in 2022. Four of its five categories require no incivility at all — persistence alone will do.

Misconception
“We can’t tell you, for privacy reasons” is usually not the Privacy Act.

Section 7B(3) exempts private-sector employee records from the Australian Privacy Principles entirely. The real constraints are defamation risk and internal policy.

Read the full report →
Report 02 25 July 2026 · Companion to Report 01
Working research

What the law was for

Australia’s worker-protection legislation of the last fifteen years, judged against what its own authors said it would achieve. The intent is documented, coherent and defensible. The outcome is largely unmeasured — and where it can be measured, it has mostly not arrived.

Findings
Holds up
The intent was honestly stated.

The Boland Review and Respect@Work were evidence-led and aimed at prevention. Nothing in the record supports a cynical reading of why these duties exist.

Gap in record
No outcome measure was ever built.

No quantified predicted outcome appears anywhere in the national chain, and the ten-year national WHS strategy sets six numeric targets — none of them for psychological injury, the fastest-deteriorating measure in the country.

Gap in record
Victoria deleted the one dataset that mattered.

The six-monthly reporting of aggression complaints recommended by its own impact statement was dropped before commencement — the only mechanism in Australia that would have produced a longitudinal record of third-party aggression against workers.

Mechanism
A prevention duty with no specified outcome becomes a documentation duty.

Where the only observable output is a policy, a register and a training record, those are what get produced — and the cheapest visibly-protective control is restricting the other party, who never appeared in the impact analysis at all.

Read the full report →
Also from the team

The research index

Shorter pieces, kept separately from the long-form reports.

· What organisations are actually required to record about a complaint.
· Our pre-registered protocol for measuring conversational intake.
· How we publish complaint trends while holding nothing about any person.
· The early-warning position paper.
Open the research index →
Corrections and contact

If we have something wrong, tell us.

We will correct it in place and date the change.
Published by The Trumis team, under our own name
Bylines None — and no endorsement implied
Acknowledgement Within two business days
Contact the team →