Answer extracted from the Beyond The Plan podcast — listen to the full episode below.
How did confusion between governance and policing play into RoboDebt's punitive outcomes?
The RoboDebt program collapsed because governance structures and policing actions became indistinguishable—what should have been oversight became punishment. When the program team executed governance through technology and process in ways that felt punitive to citizens, the line between legitimate control and harassment vanished, cementing a reputation that persists years later.
The distinction between governance and policing is subtle but critical. Governance exists to ensure the right people make informed decisions about risk at the right time. When executed properly, it creates clarity and trust. But in RoboDebt, the mechanisms meant to oversee and manage the program transformed into something far more aggressive—citizens experienced governance actions as direct punitive enforcement against them personally.
This shift happened not because anyone explicitly decided to punish people, but because the program's technology and process design blurred the boundary. As Isabel Gray-Garraway explains in the episode, the stakeholders and program team believed genuinely in the governance structure they had created. The intent was sound. Yet the execution—the algorithms, the automated letters, the debt recovery mechanisms—landed on citizens as coercive rather than corrective.
Isabel Gray-Garraway — Director of ICT Corporate Portfolio at Fire and Rescue NSW, with 25 years of project delivery experience across defence, government, and public-private hybrid organisations. She spent 17 years in government roles learning the relationship between risk, accountability, and governance. She led a 100-person team managing a $200 million rolling program at Transport for New South Wales.
The consequences were profound. The Australian public remembers RoboDebt not as a flawed governance system but as a punitive program targeting vulnerable people. This perception gap—between what was intended as oversight and what was experienced as persecution—damaged public trust in government institutions far beyond the program itself. The Royal Commission held three years after RoboDebt ended found the governance failures so severe that the government ultimately spent close to $2 billion on refunds, repayments, reimbursements, and compensation.
What makes this case instructive is that the confusion between governance and policing was not inevitable. It emerged from how governance was operationalized—from the technical systems chosen, the process flows designed, and the absence of human judgment at critical intervention points. The full episode explores how organizations can redesign governance structures to remain oversight-focused rather than drifting into enforcement, and why this distinction matters for both project outcomes and public confidence.
The lesson extends beyond RoboDebt. Whenever governance becomes rigid, automated, and disconnected from human context, it risks being experienced as policing. The difference lies in whether governance protects the program and its stakeholders, or whether it becomes a mechanism that inadvertently harms the people the program is supposed to serve.
See also
How does poor governance experience transfer across organizations and impact future project outcomes?
Negative governance experiences are transferable across time and organizations. When people move to a new organization, they carry learned behaviors and expectations shaped by past governance failures.
Why does good governance often have a poor reputation while bad governance is highly memorable?
Good governance is invisible because it works seamlessly—processes flow together, decisions are made, project teams feel supported, and it becomes business as usual. Bad governance is unforgettable because of the disruption and pain it inflicts.
What financial and reputational consequences has the Australian government faced as a result of RoboDebt's governance failures?
The Australian government spent close to $2 billion on refunds, repayments, reimbursements, and compensation, including class action payments. Three years after the program ended, reputational damage remains severe.