The Process Is the Filter.

The Saskatchewan Health Authority began work on the Administrative Information Management System with an initial budget of approximately eighty-six million dollars. AIMS was scheduled to go live in May 2021. The launch slipped to November 2022. The November 2022 launch was aborted "after a few days," in the public characterisation of the president of the Saskatchewan Union of Nurses, due to widespread functionality errors.10 The system was reintroduced in phases starting June 2024. The first payroll run under the reintroduced system, on 5 July 2024, produced documented missed and incomplete payments. A scheduling phase was rolled out in September 2025. By October 2025, the scheduling phase had caused enough operational damage — workers unable to pick up shifts, healthcare facilities understaffed, pay errors continuing — that the Province of Saskatchewan ordered the Saskatchewan Health Authority to cancel the AIMS scheduling component and revert to the legacy system it was meant to replace.12 The reversion is to be complete by 30 November 2025.

The Provincial Auditor of Saskatchewan had projected the total at two hundred and forty million dollars in 2023 — nearly triple the initial budget.2 The system's payroll, supply-chain, finance, and human-resources components — used by approximately fifty thousand employees — remain in use, with the Saskatchewan Health Authority stating they "remain stable."11 The scheduling component, the one most visible to frontline workers and most directly affecting their ability to be paid for the hours they work, has been removed from service.

This is what a quarter-billion-dollar project looks like when it does not work.

The public record on the human impact of AIMS is extensive and sourced to four major healthcare unions: the Canadian Union of Public Employees (CUPE), the Service Employees International Union West (SEIU-West), the Saskatchewan Union of Nurses (SUN), and the Saskatchewan Government and General Employees Union (SGEU) verified.9

Healthcare workers have reported, through their unions, on the public record, the following: not being paid for periods of up to five weeks; being paid less than they were owed; being paid at incorrect rates; physician invoices delayed during transition; inability to pick up scheduled shifts because the system would not register them; missed shifts producing staffing gaps in care facilities; and the cumulative financial pressure of these failures on workers who, in the words of one union president, are "unable to pay their bills, unable to make ends meet, stressed out that they might lose their cars, even their homes."

The institutional response across the rollout period followed a stable pattern. The Saskatchewan Health Authority initially characterised the rollout as "a success overall." After widespread reporting and union mobilisation, the SHA's CEO Andrew Will issued a public apology in August 2024 acknowledging "impacts" on staff.11 After the scheduling-phase failures became impossible to absorb in late October 2025, the responsible minister Jeremy Cockrill acknowledged that the errors were "unacceptable."14 The arc, in public-language terms, is success → impacts → unacceptable — and each step was a response to mounting evidence the institution had been told about earlier and chosen to characterise differently. The information had been available the whole time. The characterisation lagged the information.

The information had been available the whole time. The characterisation lagged the information.

When asked who is responsible for AIMS, the institutional answer routes the question through at least three doors.

Three institutional homes. Each can, with technical accuracy, indicate that the failure is not entirely theirs. The user community owns the consequences. The shared-services organisation owns the execution. The procurement ministry owns the central monitoring framework. No single institution owns the project end-to-end, and the institutional architecture is what permits the project, after three failed rollouts and a quarter-billion dollars, to produce no terminal accountability for any specific senior figure. Responsibility diffused across three doors is responsibility nowhere.

This is the accountability layer of the laundering. Modernisation projects of this scale in jurisdictions with single-accountable-party structures produce, when they fail, the departure of the senior figure responsible. Modernisation projects routed through three-way joint structures produce ministerial statements that the errors are "unacceptable," followed by apology, followed by the project continuing in its remaining usable components, with the senior figures remaining in post. The architecture made this outcome available; the institutions used the architecture as it was designed to be used.

In December 2025, six weeks after the scheduling-phase cancellation, the Provincial Auditor of Saskatchewan released the 2025 Report — Volume 2. The Report's chapter on "SaskBuilds and Procurement — Monitoring the Implementation of IT Projects" examined how the central oversight ministry monitors IT projects across most of the Government of Saskatchewan.1

The findings: between April 2024 and March 2025, the ministry monitored twenty IT projects with a combined forecast of approximately forty million dollars. Eight of those projects ended over budget. Twelve were delivered late. The Auditor identified one named example in which a client spent more than four million dollars on a project before it was cancelled "due to the vendor's inability to meet the client's requirements." Five recommendations were issued, including a finding that the ministry did not assess completeness of client-reported risks or consider the appropriateness of clients' risk-mitigation plans.

The same volume reported, in passing, that the Enterprise Business Modernization Project — a separate, parallel government-wide back-office IT initiative — was projected at two hundred and sixty million dollars to fully implement, following a one-hundred-and-thirty-million-dollar budget overrun and a two-and-a-half-year delay as of 31 March 2025.

AIMS is approximately two hundred and fifty million dollars and has just had its scheduling component cancelled. EBMP is approximately two hundred and sixty million dollars projected and is two and a half years late. The two projects together — EBMP-AIMS, as Saskatchewan editorial reporter Tammy Robert has paired them18 — are approximately half a billion dollars of attempted back-office IT modernisation, both running through the same provincial procurement architecture, both producing the same outcome at different speeds. The Auditor's broader finding — that of twenty monitored IT projects in a single fiscal year, eight ran over budget and twelve ran late — confirms that EBMP-AIMS is not anomalous. It is the headline-grade instance of a pattern operating at every level of provincial IT-project delivery.

The question this raises is the one the modernisation vocabulary is constructed to defer: if a single fiscal year's worth of centrally-monitored IT projects produces a sixty percent over-budget rate and a sixty percent late-delivery rate, the issue is not "this particular project's challenges." The issue is the capability of the institution to execute IT projects at all.

A capable institution does not produce eight-of-twenty IT projects over budget in a single year. A capable institution does not run a quarter-billion-dollar back-office system through three failed launches across three years, with healthcare workers unpaid in the interim. A capable institution does not, when asked, route the question of accountability across three institutional homes and produce no terminal consequences for any senior figure.

These outcomes are not random and they are not bad luck. They are produced by the institutional capability that exists in the rooms where the decisions are made. And institutional capability is, downstream of one thing: who is in the room. Who is in the room is downstream of one thing: who was hired into the room, who stayed, and who left.

Saskatchewan public institutions — and the provincial Crown corporations and provincial-government ministries adjacent to them — have spent two decades running hiring processes that filter for tolerance of institutional process rather than for capability to deliver. The procedural treatment a senior architect receives in a Saskatchewan public-sector competition is, in many cases, indistinguishable from the procedural treatment a co-op student receives. The forms are the same. The competency-by-numbers scoring is the same. The interview panels are the same. The presumption that the process is neutral and the candidate must absorb whatever the process delivers is the same. The process is the filter, and what it filters out is the candidate with the option to walk away — which is, by definition, the senior, capable, mobile candidate who is most needed in the room.

The process is the filter, and what it filters out is the candidate with the option to walk away.

Institutions that run these processes for two decades end up with staffing populations selected for tolerance of the process. The people who remain are by definition the people who, faced with the same procedural treatment given to junior candidates, did not walk. Some of these people are highly capable. Many are not. The aggregate effect on the institution's ability to execute complex projects — the kind of projects that require senior judgment, vendor-management capability, the willingness to push back on a procurement decision that is not going to work — is corrosive at scale and invisible turn by turn. The institution does not know what it has filtered out, because the filtered-out candidates do not show up in its reports.

AIMS is what an institution staffed in this way produces when it attempts a back-office modernisation. EBMP is what the same staffing pattern produces in the adjacent provincial ministry. The Provincial Auditor's twenty-projects-in-a-year finding is what the same staffing pattern produces in aggregate. The hiring architecture is the upstream cause. The modernisation catastrophes are the downstream symptoms. The public account treats the symptoms as the problem and the cause as a separate question. They are not separate. They are the same operation.

The hiring-architecture argument advanced in the preceding section operates at the level of structural inference — the catastrophic IT rollout is evidence that capable senior staff are not in the decision rooms, because the rooms where it was decided to proceed contained no one who could foresee the predictable failure. The argument is sound but it is, at that level, inferential. The Provincial Auditor of Saskatchewan has examined the same hiring-and-retention architecture at a different level of the same institution and found exactly the pattern the inference predicts.

In December 2022, the Provincial Auditor published Chapter 12 of the 2022 Report — Volume 2: "SHA Needs Targeted Plans to Fill and Retain Hard-to-Recruit Healthcare Positions."19 The findings were not subtle. The Saskatchewan Health Authority identified thirty-one positions as hard-to-recruit. The overall vacancy rate as of 31 March 2022 was 5.1 per cent — approximately two thousand unfilled positions across the institution. The five-year forecast shortage was two thousand two hundred workers across the hard-to-recruit categories, including eight hundred and forty continuing care assistants, five hundred and twenty registered nurses, and one hundred and eighty medical laboratory technicians.21 The Auditor's own characterisation of the consequence, on the public record: emergency rooms and lab services were unavailable to Saskatchewan residents because the institution did not have the staff to deliver them.

Seven recommendations were issued. Three years later, in December 2025, the Provincial Auditor published Chapter 24 of the 2025 Report — Volume 2: the follow-up review of those recommendations.20 The finding: three of the seven had been fully implemented. Four remained outstanding. The four outstanding recommendations are the structurally significant ones. The SHA had not yet assessed which facility locations were most likely to have shortages. It had not considered the root causes of those shortages. It had not developed targets for performance measures in its First Nations and Métis recruitment and retention plan. It had not analysed the results of staff exit surveys to inform retention strategies. The Auditor's chronic-vacancy-rate measure was close to or above the SHA's own 5 per cent target in June 2022, and June 2023, and June 2024 — three consecutive years of failure to hit a target the institution had set for itself.

Provincial Auditor Tara Clemett characterised the situation directly in December 2022: "They're not figuring out why people aren't staying." And, on the rural-and-remote dimension: "What do they need to do to get people to want to work in rural and remote Saskatchewan?" The questions, three years later, are still open. The exit-survey analysis that would identify the root causes of staff departure — the analysis that, if conducted, would tell the institution why it cannot retain the people it hires — remains, as of the Auditor's December 2025 finding, unimplemented.

The exit-survey analysis that would identify why staff leave remains, three years after the Auditor recommended it, unimplemented.

The pattern is the same one identified at the senior-IT-capability level in § 05, with the same shape and the same consequence at the workforce level the Auditor can directly audit. The institution can hire — the SHA's CEO reported in December 2025 a 56 per cent reduction in chronic nursing vacancies between September 2023 and September 2025, and a 64 per cent reduction in rural and northern chronic nursing vacancies in the same period, attributing the improvement to multiple programs including new-graduate training and recruitment of nurses from outside Canada.25 The numbers are real. The improvement on the single metric is real. But the understanding of why people leave — which the Auditor first recommended in 2022 and which the SHA had not implemented by August 2025 — is not a number on a dashboard. It is the structural question the dashboard cannot answer. An institution that improves its hiring rate without ever determining why it loses the people it hires is, on the public record, running a recruitment apparatus disconnected from its retention apparatus. New hires arrive; the institution does not ask why prior hires left; the same conditions that produced the departures continue to operate; the new hires, in time, depart for the same reasons. The annual cycle resets. The replacement rate is the metric. The retention question is the one the institution has, by its own documented behaviour over three years, declined to ask.

This is the architectural pattern made directly visible by the Auditor at the level of the institution's own bedside workforce. At the senior-IT-capability level, the same architecture is necessarily inferred from outcomes — the failed rollouts, the cost overruns, the diffused accountability — because the senior-IT-capability hiring decisions are not directly audited. But the same institutional architecture makes both decisions. The institution that, after three years, has not asked why its registered nurses leave is the same institution that, on the IT-modernisation file, did not ask whether the AIMS vendor relationship was producing the capability the institution needed. The same management posture that treats recruitment metrics as proof of strategic success on the front-line workforce treats project-launched milestones as proof of strategic success on the IT-modernisation file. Both responses launder a deeper question: what does the institution actually understand about its own operations, and what would change if it understood more?

The Provincial Auditor of Saskatchewan can document the staffing pattern with audit precision because staffing is in the Auditor's mandate. The hiring architecture at the senior-IT-capability level is not, directly, in the audit mandate — but the structural inference from outcomes is now reinforced by the directly-audited finding at the workforce level. The two levels are the same institution running the same architecture. The Auditor's finding at the bedside level is the evidence the § 05 inference at the senior-IT level no longer needs to make on its own.

It does not claim that any specific senior official at the Saskatchewan Health Authority, 3sHealth, or SaskBuilds and Procurement acted in bad faith. The structural argument does not require the bad-faith claim. The institution's culture is what produced the outcome; the individuals within it are the institution's products, not its authors.

It does not claim that the 3sHealth or SHA workforce as a whole is incompetent. The systems that run, when they run, are run by people who deserve credit for the parts that work. The failure is the institution's, not the staff's; the staff are the ones absorbing the cost of the failure most directly through the very pay errors the case documents.

It does not claim that AIMS or EBMP can be specifically traced to one identifiable hiring decision. The hiring-architecture argument operates at the level of culture sustained over two decades, not at the level of a single procurement.

It does not claim that vendors involved in AIMS are the principal cause. Vendor responsibility — and the broader cross-Crown pattern of vendor capture that produces the demand for projects of this kind — is the subject of Case 11 (The Legacy Annuity), forthcoming in this volume.

It claims this: the Saskatchewan Health Authority's AIMS catastrophe is one instance of a recurring pattern in Saskatchewan public-institution IT delivery; that pattern is documented by the Provincial Auditor at the level of aggregate project-monitoring failure; the EBMP parallel demonstrates the pattern is not contained to SHA; the underlying cause is not a sequence of unfortunate technical decisions but the institutional capability available in the rooms where the decisions were made; that institutional capability is produced by the hiring architecture; the hiring architecture systematically selects for tolerance of institutional process rather than for capability to deliver; the modernisation vocabulary launders the staffing root cause; and the people paying the cost of the laundering — directly, in withheld wages — are the institution's own workers.

The strongest version of this case is the version that refuses to indict individuals and refuses to soften the institutional indictment.

The pattern is the case. The pattern is sourced. The institutional architecture that produces the pattern is observable in the public record. The choice to continue operating the architecture, after the pattern has been this thoroughly documented, is the choice the responsible parties are making.

The Provincial Auditor's audit of the SHA's own bedside workforce, in § 05.5, shows the pattern operating with audit precision at one level of the institution; the AIMS rollout shows the same pattern operating, with audit-comparable precision, at the senior-IT-capability level the audit function cannot directly reach. The architecture is one architecture.

AIMS launched in November 2022 and was aborted in days.
AIMS relaunched in July 2024 and produced documented pay failures.
AIMS relaunched again in September 2025 and produced documented scheduling failures.
The Province ordered the scheduling component cancelled on 29 October 2025.
Cumulative spending is approximately $250 million.
The Provincial Auditor's 2025 Report finds the same pattern at aggregate scale across provincial IT-project monitoring.
The Enterprise Business Modernization Project, projected at ~$260 million, runs the same pattern in a parallel ministry.
All seven statements are true.
Their relationship is the case.