Generative Artificial Intelligence (AI) is spreading faster than earlier transformative technologies, driven by low-cost, task-specific tools and ease of use, said TD.
Rapid adoption does not guarantee sustained productivity gains, writes the bank in a note to clients. As in past technology cycles, lasting gains require complementary investment in data, systems, skills, and organizational changes that embed AI into core workflows.
AI use is spreading rapidly, but integration remains uneven in Canada, states TD. The technology has the potential to lift Canadian productivity, but history shows that the realization of productivity gains depends on complementary investment and organizational change.
For AI, the ability to verify and use outputs reliably at scale is a central constraint, pointed out the bank.
Canada's weaker investment in intangibles, smaller firm size and fragmented markets raise the bar for success relative to peer economies, adds TD. Closing this gap requires that firms move beyond experimentation and build the capabilities needed for system-level integration.
This includes improvements in data connectivity, interoperability across systems and production environments that allow output to be tested, validated, and used with confidence. Where firms operate at sufficient scale and face incentives to invest in these capabilities, AI is more likely to be integrated into core workflows rather than applied at the margins, according to the bank.
The key test for Canada is whether adoption is accompanied by a shift in how production is carried out across firms. Without that shift, AI use may remain broad but shallow, improving efficiency in isolated tasks without delivering a sustained, economy-wide lift in productivity.
The strongest gains lie in the later stage of this process, once earlier investments translate into new ways of working that can be deployed at scale across the economy, concludes TD.