Employee Fails PIP, Then Gets Caught Using AI in His Unfair Dismissal Case
He Denied Using AI in His Unfair Dismissal Case. The Commission Didn't Buy It.
Most employers hope they never have to find out whether their Performance Improvement Plan will hold up in front of the Fair Work Commission. GPC Asia Pacific Pty Ltd found out, and the answer was yes.
In Koonjah v GPC Asia Pacific Pty Ltd [2026] FWC 2680, Deputy President Lake dismissed an unfair dismissal application from a former Supply Chain Planner, finding the company had a valid reason to terminate his employment and had run the process fairly from start to finish. The decision is a useful case study for any employer managing underperformance, and it comes with an unusual twist involving generative AI.
The Background
Koonjah worked as a Supply Chain Planner for GPC Asia Pacific, a distributor of automotive aftermarket and industrial components. His job involved managing a set of vendors and keeping the underlying stock data accurate and current.
His 2024 annual review rated his performance as "Needs Improvement." His manager followed up with training sessions. When performance did not lift, GPC placed him on a formal Performance Improvement Plan in April 2025, targeting four areas: meeting deadlines, seeking clarification on tasks, understanding his core responsibilities, and communication.
The PIP ran for twelve weeks, with scheduled review meetings roughly every week. Two written warnings followed when improvement did not stick. By August 2025, suppliers were flagging data errors traced back to Koonjah's work, and GPC issued a show cause letter followed by termination.
What the Commission Found
Deputy President Lake worked through each factor in section 387 of the Fair Work Act and found in favour of the employer on nearly every point:
Valid reason: The Commission found the performance issues were real, well documented, and directly linked to business risk, including stock shortages caused by outdated order data.
Notice and opportunity to respond: Koonjah had been on notice since his January review, received two written warnings, and was given a formal show cause opportunity before termination.
No bias: Koonjah alleged his manager was biased against him. The Commission found no evidence to support that claim.
No harshness: Long service, income loss, and a request for redeployment were all considered and did not tip the balance. The Commission noted GPC was under no obligation to redeploy him, and that moving departments would not have solved the underlying performance gaps.
The result: the dismissal was not harsh, unjust, or unreasonable. Application dismissed.
The AI Twist
The most eyebrow raising part of the decision has nothing to do with stock data. The Commission's directions asked Koonjah to disclose whether he had used generative AI to prepare his materials. He categorically denied it, telling the Commission his submissions were prepared in consultation with people experienced in workplace relations matters.
The Deputy President was not convinced. The submissions cited real cases for propositions those cases do not actually support, a pattern the Commission described as a hallmark of AI hallucination. One citation link even pointed to the wrong case entirely. The Commission's assessment was blunt: unless the people Koonjah consulted were named Claude and Gemini, the denial was hard to believe.
That finding did real damage. The Commission noted that Koonjah's credibility was "not improved" by the apparent decision to use AI and then deny it under a direct question from the Commission.
Why This Case Matters for Employers
A few practical takeaways stand out:
A properly run PIP is strong evidence. GPC's documented weekly check ins, clear performance criteria, and written warnings gave the Commission a clear record to assess. Vague or undocumented performance management is much harder to defend.
Give employees a genuine chance to respond. GPC's show cause process, including a written deadline and a follow up meeting, satisfied the Commission that Koonjah had a fair opportunity to be heard before the final decision was made.
Redeployment is not mandatory. Employers are not automatically required to consider moving an underperforming employee to another role before dismissing them, particularly where the performance issues would follow them anywhere.
Evidence discipline matters, even when you win. The Commission criticised GPC for withholding some of its weekly review documentation until the day of the hearing, reducing the weight given to it. A win does not mean the process was flawless.
AI generated submissions carry real risk. Litigants who lean on AI tools to draft submissions without checking the underlying cases risk citing authority that says the opposite of what they claim. Denying it when asked directly compounds the problem.
The Bottom Line
GPC Asia Pacific's outcome was not the result of luck. It reflected months of documented feedback, clear expectations, and a process that gave the employee every reasonable chance to improve before termination became the only option. For employers navigating a similar situation, this decision is a solid template for what a defensible performance process looks like in practice.
This article discusses a published Fair Work Commission decision and is provided for general information only. It is not legal advice. Employers dealing with a specific performance management or termination issue should seek advice from a qualified employment lawyer.