CIBC is piloting a more advanced artificial intelligence tool among a select group of employees, weighing the substantially higher licensing and computing costs against potential productivity gains as the Canadian bank continues refining its broader AI strategy.
A Careful, Cost-Conscious Approach to AI
CIBC has built its reputation in the Canadian banking sector around a deliberate, enterprise-wide approach to AI adoption, prioritizing governance and measured rollout over flashy, high-profile applications. That philosophy appears to be shaping how the bank is evaluating its newest and more powerful AI tool, with executives weighing whether the added expense of more capable models translates into commensurate productivity improvements before expanding access more broadly.
Richard Jardim, CIBC’s Senior Executive Vice-President and Chief Technology and Information Officer overseeing Global Technology, Data, and AI, has previously emphasized that cost is always a central factor in these decisions. Vendor-based AI solutions typically come with ongoing licensing and subscription fees that must be weighed against measurable productivity gains, while more customized or advanced tools often carry higher upfront costs that need to prove their value before wider deployment.
Building on an Established AI Foundation
This latest pilot builds on CIBC’s existing generative AI infrastructure, most notably its internally developed CIBC AI platform, which the bank rolled out starting with a pilot phase in July 2024 involving employees across all lines of business in Canada, the United States, and the United Kingdom. That platform, built on similar technology to ChatGPT, has since expanded to tens of thousands of knowledge workers across the bank, requiring completion of mandatory training before employees are granted access.
The bank has also equipped its developer community with GitHub Copilot, which Jardim has said delivered a 10% to 15% boost in productivity during pilot testing, helping accelerate development cycles and improve code quality. Separately, CIBC’s CRTeX platform, which uses AI to help frontline staff personalize client recommendations, has helped generate more than $1 billion in new client deposits since its national launch.
Governance Remains Central to the Strategy
Consistent with its broader approach, CIBC has maintained a structured governance framework around all of its AI deployments, reviewing individual use cases and building in safeguards to limit the risk of AI-generated errors or hallucinations. When CIBC AI processes documents or generates outputs, the system assesses its own confidence level, and any output falling short of established thresholds gets handed off to a human reviewer rather than being delivered directly to employees or clients.
That cautious posture extends particularly to revenue-generating applications of AI, which Jardim has said carry considerably more risk than internal productivity tools. Executives across CIBC have repeatedly stressed that AI-enabled platforms touching loans or other client-facing products and services require significantly more oversight than tools designed simply to summarize documents or draft internal communications.
Measurable Results So Far
CIBC’s AI investments have already delivered tangible efficiency gains. CIBC Chief Executive Harry Culham has said AI has helped the bank save 1.2 million hours of work, while also enhancing fraud detection and credit monitoring capabilities. Those time savings, Culham noted, have freed up capacity for the bank to pursue other growth opportunities rather than simply cutting costs.
The bank’s approach mirrors broader trends among Canada’s largest lenders, with rivals like TD Bank also touting significant AI-driven efficiency gains, including cutting mortgage approval reviews down from hours to minutes. Across the industry, Canadian banking executives have generally framed AI adoption as a long-term competitive necessity, even as they emphasize the importance of careful, human-led oversight throughout the rollout process.
What Comes Next
As CIBC works through its current pilot of this more powerful AI tool, the bank’s decision on whether to expand access more broadly will likely hinge on the same cost-benefit calculus that has guided its AI strategy throughout the past two years. With thousands of employees already trained and using existing AI tools, and measurable productivity gains already documented across multiple business lines, CIBC’s leadership appears intent on ensuring any further investment in more advanced, costlier AI capabilities delivers returns substantial enough to justify the expense before committing to a full-scale rollout.
