Public Funding Momentum Meets the Challenge of Local Execution
The Government of Canada recently announced, through the economic development agency PrairiesCan, an investment of over $10.2 million to support the adoption of artificial intelligence in SMEs. This initiative is part of a broader trend: governments are increasing funding packages to encourage businesses, particularly in regional areas, to integrate these technologies to boost their productivity and overall competitiveness.
However, injecting capital only solves part of the equation. According to a study published by the Business Development Bank of Canada (BDC), although interest in artificial intelligence is growing, the majority of small and medium-sized enterprises face a major obstacle: the lack of in-house technical skills. Recruiting machine learning engineers or specialized developers is often out of reach for an SME located outside major urban centres, due to both the labour shortage and the very high salary scales in the tech sector.
The Trap of Technical Dependency and Software Debt
Faced with this skills shortage, business leaders are presented with two traditional options, each carrying major operational risks. The first is to fully outsource the development of AI tools to external consulting firms. While this approach gets projects off the ground, it creates long-term technological dependency and recurring maintenance costs that can quickly exceed the initial grants.
The second option, increasingly common with the popularization of language models, is "vibe coding." This term refers to the rapid production of applications or scripts by non-technical employees using direct prompts to consumer chatbots, without supervision or auditing. This practice raises serious concerns among security experts. The UK National Cyber Security Centre (NCSC) has warned organizations about the intolerable risks associated with the use of ungoverned AI-generated code, which can introduce critical vulnerabilities or obsolete software dependencies into information systems.
Analyses conducted by recognized security platforms reveal that nearly half of the code samples produced directly by language models without a validation framework fail basic security tests. For an SME, adopting AI can thus turn into a vector for cyber threats or a financial sinkhole of corrective maintenance.
Governed No-Code: The Alternative for Regional SMEs
It is in this context of tension between the need to innovate and technical constraints that the governed no-code approach makes perfect sense. The Quebec-based platform ProductivIA offers an environment designed specifically to allow businesses to adopt artificial intelligence without having to write a single line of code or manage complex infrastructure.
Using the Fabrique application, SME employees can describe their business needs in plain language to design custom tools. Unlike rogue development, the platform secures the process: each generated application is isolated within a secure virtual sandbox and subjected to an automated audit by internal monitoring agents before deployment. This streamlined architecture, running directly in the user's browser, eliminates the need to maintain local software dependencies and drastically reduces the attack surface for cyber threats.
The platform's central Assistant then orchestrates these custom tools. For example, a user can ask the Assistant to connect an internal document database containing company policies to an email writing tool, thereby automating customer support or administrative correspondence without requiring any manual technical integration.
Complete Sovereignty, from Software to Infrastructure
This software democratization is part of a broader sovereign vision, which is particularly relevant for businesses subject to strict regulatory frameworks such as Law 25 in Quebec. To guarantee the confidentiality of business data, the platform's orchestrator can be configured to route requests exclusively to Matania, the language model provider physically hosted in Quebec, thereby preventing any cross-border transit of sensitive data.
Furthermore, this digital transition does not require costly hardware upgrades. By pairing the application platform with the open-source operating system Boréal OS, organizations can give their existing computers a second life, preventing electronic waste and reducing the carbon footprint associated with manufacturing new machines.
Looking Ahead
The effectiveness of public innovation funding will depend on the ability of businesses to maintain and evolve the funded tools. Technological models based on governed no-code and infrastructure sovereignty offer an essential path forward to maximize regional economic benefits while protecting organizations against dependency and cybersecurity risks.