Since 2024, the Canadian government has introduced a series of funding and policy measures aimed at boosting the country's technology sectors, particularly artificial intelligence. It has also committed to addressing long-standing issues in government procurement, including through initiatives like the Defence Procurement Strategy. A critical component of this reform is the adoption of 'agile development' principles in how the government purchases technology. However, the current procurement system remains largely based on an outdated 'waterfall' approach, where nothing is tendered or built until a comprehensive specification is drafted. This approach is detrimental to Canadian innovation.
The Flaws of the Waterfall Approach
Agile development takes a contrasting approach, employing short sprints with repeated trials, tests, failures, and fixes. This method fosters continuous accountability and allows for early course corrections or termination when approaches prove ineffective. While it does not eliminate incompetence or bad actors within government, it enables both contractors and the government to learn, improve designs iteratively, and prevent small mistakes from escalating into large, costly failures—such as the ArriveCAN app debacle. Instead of relying on a single, large upfront decision, agile procurement incorporates multiple checkpoints where performance is tested and funding can be halted, thereby reducing the risk of major failures.
Waterfall requests for proposals (RFPs) typically specify the exact technology solution the government intends to purchase. A better approach would be to define the problem and allow the private sector to propose solutions. This would eliminate the need for bidders to submit extensive dossiers of technical documents, compliance matrices, project plans, corporate histories, organizational charts, and relevant experience. For large vendors, such documentation is routine and straightforward; many have dedicated teams for government tenders. However, for startups, preparing these documents consumes scarce resources, including time. Consequently, administrative requirements often determine who can compete for contracts, rather than technical performance, effectively screening out smaller firms before they can be considered.
Impact on Canadian Startups
The waterfall approach favors companies that are best equipped to produce paperwork. In practice, this often means large, established companies—many of which are foreign—that are accustomed to navigating complex procurement systems. Thus, procurement policy paradoxically works against Canadian startups and innovators. Many Canadian technology companies end up focusing on international markets because, ironically, selling to foreign governments is easier. One of the authors has had more success selling to the U.K. government, the European Space Agency, and other international institutions than breaking into Ottawa's procurement system.
Recommendations for Change
A more agile approach would enable Ottawa to test possibilities and change course as needed. Instead of writing an RFP for a long-term contract with a specific technology—essentially assuming the technology will remain stable for years—the government should open the door to new innovations. In fast-moving fields like AI, agile procurement would allow Ottawa to test technologies through small trials, pilot projects, and rapid iteration to determine what works. One of the authors was involved in a technology trial in Saskatchewan where seven companies with four different technologies won contracts simultaneously. The winner of the trial, and subsequently a large contract, was the one with the best quality-to-cost ratio, measured using parameters stated in the RFP, which itself required only minimal paperwork to answer.
