Canada Is Building Digital Government One Doorway at a Time

A nurse who moves from Halifax to Calgary will prove who she is, separately, to at least four systems: a new health card, a new nursing registration before she works a single shift, a new driver’s licence, and new school registrations for her children, each with its own login. Almost nothing she verified in Nova Scotia follows her. Canadians encounter the same problem every day: separate logins, repeated verification, and systems that stop working when a service crosses a jurisdictional boundary. The systems were never designed to operate as one.

Canada is often measured against the wrong model. International rankings assess national performance, while service delivery in Canada runs through a federation. But the deeper issue is what Canada has not yet built. This is not a website problem. It is an infrastructure problem. The shared foundations that let identity, data, services, security, and accountability move safely across governments do not yet exist in coordinated form. Canada has strategies, portals, identity initiatives, cybersecurity institutions, cloud investments, and emerging sovereign compute capacity. What it does not have is an agenda for making those pieces work as a federated system.

That is the warning behind Canada’s slide in the United Nations E-Government Development Index, from third in 2010, its peak, to forty-seventh in 2024. Methodological changes account for part of the latest movement, but they do not eliminate the broader pattern: Canada has fallen behind countries that have moved further in areas such as shared identity, data exchange, integrated services, and digital access. The ranking is a signal, not a diagnosis. The diagnosis is that Canada is measured as a unified system while operating as a federated one, with delivery authority, data, funding, and accountability split across governments. That is where digital services actually break: not at the front door of a single portal, but at the seams between systems, rules, and responsibilities.

The missing piece is not another portal or strategy. It is a digital public infrastructure agenda.

As Canada pursues digital sovereignty through initiatives such as AI for All and sovereign AI investments, the question is no longer whether governments will digitize services, but whether they can do so on foundations they collectively trust and govern.

Digital public infrastructure is the connective layer beneath digital government: trusted identity, secure data exchange, payments, common service components, and the cybersecurity and compute capacity beneath them. Digital government is the services, portals, and policies people touch. Digital public infrastructure is what lets those services function across departments, programs, and jurisdictions. The first is the storefront. The second is the road network.

The concept is already established. The OECD defines digital public infrastructure as shared systems such as digital identity, payments, and data exchange.

The Canada School of Public Service has begun describing it as the foundation for how Canadians experience public services. What is missing is not the concept. It is a deliberate agenda to build shared, reusable, and governed foundations that span jurisdictions. That agenda does not require one national system. It requires a small number of capabilities different governments can trust and reuse: recognised identity, lawful and auditable data exchange, cybersecurity that travels with shared components, and accountability that stays clear when a service crosses a boundary.

What Canada Has and What Is Missing

Canada does not start from zero. The federal Policy on Service and Digital and Digital Ambition provides direction, provinces and territories have their own strategies and platforms, and identity standards and a national cybersecurity centre already exist. The Canadian Digital Exchange Platform was developed as a federal interoperability foundation, but it is not yet a jointly governed exchange layer operating across the federation.

The government’s own reporting makes the broader gap measurable: in 2022–23, 50.6 percent of government interactions were completed online, but only 22.5 percent of services were available entirely online from start to finish. The problem is not an absence of effort. It is that these pieces were assembled jurisdiction by jurisdiction and program by program, rather than designed to interlock across the federation. Standards exist, but they are rarely connected to who pays, who decides, and who is accountable when a service crosses a boundary.

Budget 2025 announced a new Office of Digital Transformation to lead the adoption of AI and other technologies across the federal government, including made-in-Canada sovereign AI tools. Released in June 2026, Canada’s national AI strategy, AI for All, gives the office responsibility for accelerating the procurement and delivery of AI solutions across the federal government. The test is whether the office receives the authority, resources, and continuity needed to govern shared foundations while working effectively across jurisdictions. The lesson from elsewhere is blunt: digital centres without staying power produce frameworks without the conditions to apply them. Similar debates are unfolding in Australia, where observers argue that digital government still lacks the connected capability needed to translate frameworks into delivery.

Generative AI raises the stakes. Without that connective layer, Canada will repeat the first generation of e-government: instead of too many portals and accounts, too many separate chatbots, copilots, and agents, each built around the boundaries of one organisation. Agentic AI makes the duplication more consequential because these systems can initiate tasks, move data, and shape service decisions. Layered onto a fragmented state, AI is just another front end on the same broken seams.

The routine act of signing in makes that fragmentation visible. The federal government maintains a page to help people find the right account; the Auditor General counts almost 90 separate sign-in portals run by individual federal departments. It is useful, and it is telling. A directory of doorways is what you build when you cannot yet remove doorways. The Canada Revenue Agency illustrates both progress and its limit. CRA has simplified access through a single sign-in, but that is not the same as a coherent account architecture across governments. 

Identity is where this bites hardest. Canada does not need one uniform credential imposed across the country. It needs interoperable identity that reduces repeated verification and lets credentials travel across jurisdictions. In its 2024 audit, the Auditor General found that digital access at all levels of government is evolving without common parameters to ensure the interoperability of secure and reliable identity validation systems. The Pan-Canadian Trust Framework is a serious anchor, but a voluntary framework does not create the authority, incentives, or operating conditions people meet at the counter. Comparable federations are moving: Australia’s Digital ID Act 2024 creates the legislative and accreditation framework through which Commonwealth, state, and territory entities can participate in an interoperable digital identity system, directly addressing the kind of gap Canada’s Auditor General has identified.

None of this is an argument for centralisation. Estonia is useful for one principle, not as a template: it treats secure data exchange as public infrastructure rather than a problem each service solves alone. Its X-Road layer allows participating public- and private-sector systems to exchange data through a common, secure, and logged interoperability model. Canada already has emerging foundations, including the federal Canadian Digital Exchange Platform. Québec has also joined the Nordic Institute for Interoperability Solutions as a partner and has access to its open-source X-Road technology as part of its digital-sovereignty work. The point is not to copy Estonia. It is to establish common trust, security, and accountability arrangements so that every jurisdiction is not forced to solve data exchange independently.

Risk, Accountability, and Sovereignty

The logic of shared, governed infrastructure extends well beyond service delivery. Once data exchange, identity, and automation shape how governments make decisions, shared infrastructure becomes a question of risk, accountability, and sovereignty.

Australia’s Robodebt program is the warning. Automated income averaging ran across data that was not reconciled, auditable, or governed as shared infrastructure, and the result was a welfare debt scheme a Royal Commission found neither fair nor legal. Robodebt was not even AI. It was simple automation. If weak data foundations produced unlawful outcomes at scale with basic tools, the risks multiply as governments adopt systems that are faster, less transparent, and harder to audit after the fact.

Canada now has clearer national direction. AI for All and the Directive on Automated Decision-Making set meaningful requirements, but AI is already reshaping provincial, municipal, health, education, justice, and social service settings beyond federal reach. The strategy itself identifies the dependency, observing that “AI is only as powerful as the data it can access” and that the deepest data assets in the country have been built up over decades by federal, provincial, and territorial institutions. A national strategy can set direction. It cannot execute at the scale it implies without the foundations beneath it. Cybersecurity exposes the same seam: a weak point in one corner of a shared ecosystem becomes everyone’s exposure.

Seen this way, digital sovereignty is not only about foreign platforms and cross-border data. It is about internal capability: whether Canadian governments can control, secure, adapt, and account for the systems public administration now runs on.

Sovereign compute matters too, but only if governments retain meaningful control over the systems they depend on.

Canada has already made this kind of move once. The One Canadian Economy Act, which removes internal barriers to trade and labour mobility, did not decide that provincial differences no longer matter. It decided that some barriers were too costly to leave unmanaged. Digital barriers work the same way: fragmented identity, weak data reuse, uneven cybersecurity, supplier dependence, and unclear accountability now impose system-wide costs. They are simply harder to see than a blocked road or a stalled transmission line. If Canada can talk about one Canadian economy, it can talk about one Canadian digital public infrastructure agenda. Not one system. Not one portal. Not one credential. Shared foundations strong enough to let different governments operate coherently when people, data, risks, and responsibilities move between them.

Execution confidence has to be earned, not announced. Canada has shown it can modernise major systems: in March 2025, all 7.4 million Old Age Security recipients moved to a modern platform under Benefits Delivery Modernization, a programme the Auditor General had flagged as high risk two years earlier. Phoenix and ArriveCAN also showed that shared systems require governance and expertise that last beyond launch. Some federal benefits and tax systems have been in service since the 1960s, and modernisation funding remains limited and inflexible. None of this reflects a shortage of skilled public servants. It reflects the difficulty of delivering modern services on foundations never built for today’s expectations.

Proof, Not Promises

So the practical question is what would force change. A ranking will not. A strategy will not either. What Canada needs is proof.

By 2031, Canada should be able to point to one high-value pathway that crosses jurisdictions and say, credibly: this works across the federation. Not perfectly. Not uniformly. But well enough to prove that the pieces can align around a real public need.

The most revealing pathway is determining eligibility for a benefit and paying it correctly across governments. It runs through identity, shared data, program rules, and the ability to explain and correct decisions. Automated logic without reconciled, auditable data creates the risk of unlawful outcomes at scale. A serious agenda would make lawful data movement, explainability, and correction part of the infrastructure, not an afterthought. 

Canada would need to pick one hard problem and assign clear ownership. It would need to define federal, provincial, territorial, municipal, Indigenous, and service-delivery roles up front, decide which credentials are trusted and what data can move under what authority and safeguards, and build privacy, cybersecurity, accessibility, audit, explainability, and redress into the design before launch.

It would also need to keep enough engineering, architecture, procurement, and product capability inside government to reuse what works instead of creating another one-off solution. Shared components are what turn standards into routine practice.

A proof pathway also forces a better question about measurement. Canada should keep tracking international rankings, but pair them with a federation-specific assessment of whether systems, rules, and accountability work across boundaries. The UN’s own technical appendix acknowledges that its index may not fully capture digital progress at the state level in large federal countries, where subnational governments do much of the delivery. That is precisely Canada’s situation. Such an assessment should not flatten Canada into a single score; the provinces and territories are not starting from the same place. A federation interoperability scorecard would ask more practical questions: whether credentials can be reused across jurisdictions, whether program-to-program data exchange uses a common layer, how quickly wrong decisions can be explained and corrected, and whether shared components contain incidents before they spread.

That is the real meaning of Canada’s digital slide. Not that the country failed to modernise. Not that nothing works. Not that it needs one central system. The warning is narrower and sharper: comparable federations are beginning to legislate and govern the shared foundations that Canada has named but not yet built.

The challenge is no longer digitisation. It is integration.

Five years from now, the question should not be whether Canada has more strategies, pilots, data centres, or services online. Success would feel ordinary: verify once, update once, correct errors quickly, and move through public services without having to understand which government boundary has been crossed.