Scotiabank reported a 95% adoption rate for Scotia Navigator, its AI assistant launched in April to support employees with research, analytics and coding tasks.
Employees at the Toronto-based bank used AI 17% more to help answer customer questions in its fiscal third quarter, which ended July 31, compared with Q2, Scotiabank reported. Also, more than 80% of the bank’s employees have completed at least one AI course, according to today’s earnings presentation.
Scotiabank continues to invest in AI “to advance our enterprisewide AI agenda with a focus on practical adoption including training, scalable infrastructure and responsible governance,” President and Chief Executive Scott Thomson said on today’s earnings call.
In April, Scotiabank introduced Scotia Intelligence, an initiative to accelerate its AI adoption and equip employees with secure data and AI tools. The bank expanded Scotia Intelligence in Q3, adding capabilities to improve employee productivity and workflow, Thomson said.
“These new advanced features will help our teams collaborate in real time, turn complex information into clear outputs and move from concept to execution faster,” Thomson said.
“With the recent launch of our Scotia Intelligence knowledge agents, employees now have access to AI-powered solutions that facilitate easy access to institutional information, enabling faster execution of routine processes, helping them to focus on higher value innovation and client outcomes,” he said.
AI Consortium
Furthering Scotiabank’s commitment to AI implementation, the bank co-launched the AI Consortium, “a collaborative Canadian model designed to help large, regulated organizations build and govern the critical control systems required to deploy AI safely,” Thomson said.
It was launched with technology firm Lightworks, financial services company Sun Life and communications technology company Telus in July.
The consortium’s flagship program, the Agentic Control Plane, helps enterprises manage agentic AI at scale, a July 7 release said. Future projects include:
- An AI operations center to help members improve AI performance, resilience and cost management; and
- A secure framework for exchanging AI tokens.
Evolving cyberthreats
AI advancements are enabling new types of cyberattacks, Scotiabank noted in its quarterly report to shareholders, published today. Frontier AI models, such as Anthropic’s Claude Mythos 5, can rapidly find an institution’s previously unknown vulnerabilities, the report said.
In response, Scotiabank is transitioning to a real-time security model focusing on AI-enabled detection, automated response and prevention, and scalable remediation, the bank said. It also has cyber insurance coverage to help mitigate any potential losses linked to cyberattacks, the bank said.
By the numbers
The CA$1.5 trillion ($1.1 trillion) bank has grown its digital banking capabilities in Canada, with digital sales reaching 40% of total sales in Q3, up 840 basis points year over year and 180 basis points quarter over quarter. The bank reported a 2.9% YoY increase in active digital users to 5.3 million and a 4% YoY increase in active mobile users to 4.8 million.
Scotiabank also reported in Q3:
- $7.6 billion in total revenue, up 11.1% YoY; and
- $2.1 billion in net income, up 16.9% YoY.
Shares of Scotiabank (TSX: BNS; NYSE: BNS) were up 7.16% from market open to $93.08 as of market close today. The bank has a market capitalization of $114.2 billion.
Editor’s note: All amounts have been converted to U.S. dollars.
Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas. This inaugural event will include speakers from Fifth Third and Capital One, as well as a fireside chat with Piermont Bank founder and Chief Executive Wendy Cai-Lee.
Facts Only
* Scotia Navigator adoption rate is 95%.
* Employees used AI 17% more for customer questions in Q3 compared to Q2.
* More than 80% of bank employees have completed at least one AI course.
* Scotia Intelligence was introduced in April to accelerate AI adoption and equip employees with secure data and AI tools.
* Scotia Intelligence capabilities were expanded in Q3 to improve employee productivity and workflow.
* Knowledge agents provide access to institutional information for faster execution of routine processes.
* The bank co-launched the AI Consortium with Lightworks, Sun Life, and Telus in July.
* The Agentic Control Plane is the flagship program of the consortium, designed to manage agentic AI at scale.
* Scotiabank transitioned to a real-time security model focusing on AI-enabled detection, automated response, and scalable remediation.
* The bank reported $7.6 billion in total revenue, up 11.1% year-over-year.
* The bank reported $2.1 billion in net income, up 16.9% year-over-year.
* Digital sales reached 40% of total sales in Q3.
* Active digital users reached 5.3 million in Q3, an increase of 2.9% year-over-year.
* Active mobile users reached 4.8 million in Q3, an increase of 4% year-over-year.
* Shares of Scotiabank were up 7.16% to $93.08 at market close.
Executive Summary
Full Take
The narrative frames AI adoption not merely as a productivity tool but as a strategic imperative tied directly to institutional governance and security, which demands a structured approach beyond simple tool deployment. The simultaneous push for practical employee upskilling (courses, Navigator use) and the creation of a formal regulatory/governance structure (AI Consortium, Agentic Control Plane) suggests an attempt to manage the inherent risk of rapid innovation through formalized control systems. The focus on "practical adoption" alongside "responsible governance" positions AI implementation as a balancing act between agility and systemic safety.
The pattern observed is one of proactive risk mitigation integrated into the innovation cycle, rather than reactive compliance after deployment. The juxtaposition of accelerating knowledge access (Scotia Intelligence) with establishing external frameworks for safe deployment (AI Consortium) suggests an awareness that capability without control introduces unacceptable exposure. The mention of frontier models finding unknown vulnerabilities highlights a specific fear appeal: that the very advancement being pursued is inherently destabilizing. This leads to a systemic framing where AI success requires institutional maturity, pushing beyond mere feature adoption toward creating secure operational and ethical boundaries.
The central tension lies between the velocity of technological capability and the necessary slowness of establishing responsible control mechanisms. The implication for agency is whether these large, regulated entities can successfully embed governance structures that scale effectively across distributed, evolving AI agents without stifling the very innovation they seek to leverage. What systems are in place to ensure that agentic capabilities remain aligned with institutional objectives rather than devolving into unmanaged execution?
Sentinel — Human
The text reads like a factual summary of a corporate earnings report, exhibiting the predictable structure of financial news rather than purely synthetic prose.