Investigations

Before any show cause letter goes out, someone has to work out what actually happened. How well that investigation is run — who does it, how open-minded they are, and how much evidence they actually gather — shapes everything that follows.

What you’ll be able to do

• Explain what makes an investigation fair, independent and genuinely open-minded, as distinct from whether notification and the opportunity to respond were timed correctly.
• Apply the right standard of proof to disputed facts, and know when Briginshaw demands more care.
• Spot the point where relying on someone else’s report, rather than first-hand evidence, becomes a risk rather than a convenience.

How this fits with the rest of s.387

This module is about the investigation itself — not when the employee was told, or whether they got a chance to respond (that’s Module 6, on s.387(b)/(c)). An investigation can be run well or badly well before any show cause letter exists. Its quality feeds into whether the Commission accepts the employer’s findings of fact under s.387(a), and — where it doesn’t fit squarely under ss.387(b)-(e) — is weighed as a procedural matter under the s.387(h) catch-all.

The investigator must be free of a real or apparent conflict

An enterprise agreement commonly requires those involved in a disciplinary process to act “without favouritism, bias or conflict of interest” and to “act fairly and impartially.” Even without such a clause, the Commission scrutinises whether that happened in substance.

In Haslam v SAE Institute, a lecturer was dismissed after a student complaint. The investigator’s findings on one allegation changed from unsubstantiated to substantiated — and the recommendation from a warning to termination — only after the decision-maker pushed back and the complainant’s mother applied pressure. The employee argued the decision-maker was “reverse engineering the outcome he was seeking” [2025] FWC 3826 at [52], pointing also to the decision-maker’s own language that the employer was “taking a victim-centric approach” [2025] FWC 3826 at [53]. The Commission found the decision-maker “did not make a proper or fair assessment of either the evidence in relation to allegation 2 or the objective seriousness” of the other allegation [2025] FWC 3826 at [58], and ultimately found the dismissal unjust and unreasonable because the employer failed its own obligation to act fairly and impartially in the disciplinary procedure [2025] FWC 3826 at [62].

But a bare assertion of conflict isn’t enough either. In Gunawardana v Services Australia, the employee argued the investigation was tainted because one manager had been involved both in his current investigation and an earlier one. The Commission rejected that: mere dual involvement, without evidence of actual partiality, doesn’t establish a flaw [2021] FWC 2243 at [146]. A conflict claim needs to point to something the conflicted person actually did, not just their presence in the room.

Allegations need enough detail that the employee can actually answer them

This is distinct from the timing question in Module 6 — here the question is whether what was put to the employee, in substance, was specific enough to engage with. In Keron v Westpac, the allegations put to the employee at the investigation meeting identified the date, approximate time, location, the precise physical conduct alleged, and its approximate duration [2022] FWC 221 at [100]. That’s the level of specificity an investigation should be aiming for — not a category, not a generality, but something the employee can genuinely admit, deny or explain.

Don’t predetermine the outcome before hearing the employee’s side

An investigation that starts from a conclusion and works backwards to justify it isn’t an investigation. The Haslam findings above are the clearest illustration: a decision-maker who directs an investigator toward a particular result, rather than letting findings follow the evidence.

Gunawardana shows a subtler version of the same problem. The external investigator there had, by his own terms of reference, declined to inquire into whether the employee’s counter-complaints were actually substantiated or correct — treating that as “out of scope.” The Commission was unpersuaded: if investigating “unsubstantiated” complaints was the point of the exercise, it is hard to see how that could be done without inquiring into whether they could, in fact, be substantiated [2021] FWC 2243 at [140]-[141]. The investigator’s failure to properly engage with the employee’s detailed response to the draft report, before finalising it, was itself a finding against the employer [2021] FWC 2243 at [142].

An investigation that only tests the employer’s version of events, and treats the employee’s response as something to be filed rather than engaged with, has already predetermined its outcome — whatever the eventual report says.

The standard of proof: balance of probabilities, with more care for serious allegations

The employer carries the onus of proving disputed conduct happened, and it’s assessed on the balance of probabilities — not on reasonable suspicion, and not on what the employer honestly believed after its own process. In Bowen v Cape York Grassroots Aboriginal Corporation, the Commission applied the classic Briginshaw formulation directly to a misconduct investigation: “the standard of proof remains the balance of probabilities but ‘the nature of the issue necessarily affects the process by which reasonable satisfaction is attained’ and such satisfaction ‘should not be produced by inexact proofs, indefinite testimony, or indirect inferences’” [2023] FWC 1198 at [81].

In practice, that means the more serious the allegation, the more the investigation has to actually nail down — not just collect an impression. Haslam made the same point from the other direction: where serious misconduct is alleged, “the Briginshaw standard applies so that findings that an employee engaged in the misconduct alleged are not made lightly” [2025] FWC 3826 at [14].

The risk of relying solely on a report, without first-hand evidence

An investigation report is a useful record of what was found — but it is not a substitute for evidence the Commission can itself test, and treating it as such carries real risk.

In Dickman v Ventia Australia, the employer’s investigators had doubts about an employee’s candour but declined to speak to a key witness, relying instead solely on their own analysis of one email and the employee’s response to it. The Commission found that left the employer without “a sufficiently reliable counterfactual” against which to test the employee’s denial, so it was not reasonably open to find he had been dishonest [2024] FWC 2914 at [162]. No valid reason was established, and the dismissal was found unfair [2024] FWC 2914 at [229].

Haslam illustrates the same point from the Commission’s own vantage point at hearing: where an investigation produces only a written account from one side, and the employee gives live evidence, the Commission may simply prefer what it can see and test directly — “I have had the benefit of hearing and seeing Mr Haslam give evidence, whereas I only have the written accounts from Student A” [2025] FWC 3826 at [31]. A report-only investigation leaves an employer exposed exactly where a hearing turns on credibility.

A flawed investigation doesn’t always sink the case — but don’t rely on that

Importantly, the Commission doesn’t treat every investigative imperfection as fatal. In Gunawardana, despite finding real flaws in how the investigator handled the employee’s response, the Commission still found a valid reason for dismissal — because the underlying findings were independently supported by a comprehensive body of contemporaneous evidence, much of it created by the employee himself [2021] FWC 2243 at [152]. The lesson isn’t “flaws don’t matter” — it’s that a flawed investigation survives only when there’s a solid, independent evidentiary foundation underneath it. Don’t build a case that depends on that foundation existing; build the foundation.

What the Commission looks for

• Who ran it. Was the investigator free of a real conflict — not just free of the appearance one party wants to allege, but genuinely not standing to gain or lose from a particular outcome?
• Whether findings followed evidence, or the reverse. Did the investigator’s conclusions change only when new evidence came in, or when pressure was applied?
• Specificity. Could the employee actually work out, from what was put to them, what they were said to have done?
• Engagement with the response. Was the employee’s account of events actually tested against the evidence, or filed and moved past?
• The evidentiary base. Does the finding rest on witnesses, documents, footage or admissions the employer (and the Commission) can test — or on an inference from limited material that was never put to the obvious person who could confirm or deny it?

How it fails

Adviser checklist

Practice scenario: the agency team lead

Decisions cited

Check each paragraph reference against the published decision on the FWC website before relying on it. Briginshaw v Briginshaw is a 1938 High Court authority quoted within one of the decisions below and is not independently in this corpus — check the quote against the original. Gunawardana v Services Australia [2021] FWC 2243 was unsuccessfully appealed; the Full Bench refused permission to appeal, leaving the decision undisturbed.