Scotiabank is pushing deeper into artificial intelligence, saying its new "knowledge agents" are helping employees turn internal information into completed tasks. The Bank of Nova Scotia, one of Canada's largest lenders, said more than 71,000 staff can now access its Scotia Intelligence platform, and that employees have logged 14 million actions since the tool launched in March 2026.
The bank is pitching Scotia Intelligence as more than a souped-up document search. Instead of just finding a policy or a past report, the AI is designed to pull from approved internal sources, connect information across different systems, and then help staff complete routine steps—like drafting a response, filling out a form, or preparing a summary for a client meeting.
What are 'knowledge agents'?
In plain terms, a knowledge agent is an AI system that doesn't just answer questions but takes actions based on the information it finds. Think of it as a digital assistant that can read your bank's internal manuals, pull up the relevant customer history, and then generate a draft email or update a record—all within the guardrails of the bank's approved data.
For a bank, that's a big deal. Financial institutions sit on enormous amounts of data—loan files, compliance rules, market research, customer records—but much of it sits in silos. Knowledge agents aim to stitch those silos together, letting employees spend less time hunting for information and more time using it.
Scotiabank says the adoption has been broad. More than 80% of its global workforce has completed at least one internal AI training course, and that figure rises to over 90% for directors and above. That suggests the bank is trying to make AI fluency a baseline skill, not a niche specialty.
Why this matters for investors
For investors, the headline number is the 14 million actions. That's a measure of real usage, not just a pilot project. When a bank's employees are logging millions of actions on an AI platform, it's a sign that the technology is moving from experimentation to everyday work.
Banks have been investing heavily in AI for years, but the payoff has often been slow. The hope is that tools like Scotia Intelligence can cut costs, speed up service, and reduce errors—all of which can boost profitability over time. If Scotiabank can show that AI is actually improving productivity, it could give the stock a leg up compared with peers that are slower to adopt.
But there are also risks. AI systems in banking face strict regulatory scrutiny, especially around data privacy and fairness. A mistake—like an AI giving wrong advice to a customer or mishandling sensitive data—could damage trust and invite fines. Scotiabank says the platform draws only from approved internal sources, which is a nod to those concerns.
What to watch next
Investors will be watching whether the 14 million actions translate into measurable financial results. Look for mentions of AI-driven cost savings or efficiency gains in Scotiabank's quarterly earnings calls. Also watch for any expansion of the platform to customer-facing services, which could signal a bigger strategic bet.
The broader trend is clear: AI is becoming a competitive battleground for banks. ChipAgents raises $60M to speed up chip verification with AI agents shows that AI agents are also making inroads in other industries, from chip design to customer service. For banks, the question isn't whether to use AI, but how quickly and how safely they can scale it.
Scotiabank's numbers suggest it's ahead of the curve in adoption. Whether that leads to better returns for shareholders will depend on execution—and on whether the AI actually delivers the productivity gains the bank is promising.
