Scotiabank gives its Canadian commercial bankers two AI agents for credit work

Anyone who has trained as a commercial banker knows the ritual. A new deal arrives, the financial statements land as a stack of pages and the first hour goes on keying figures, calculating ratios and checking that the numbers tie. It is the apprenticeship of the craft, and for decades it has also been where much of the craft's time quietly went. The analysis was the price of admission to the client conversation, and it often left less room for the conversation itself.
Scotiabank is now betting that this hour can shrink to almost nothing. Its Commercial Navigator, available to the entire Canadian Commercial Banking team, gives employees two agents. The Financial Statement Analysis Agent speeds initial financial analysis on new deals from one hour to less than one minute, returning key ratios, trends and questions to raise with clients. The Credit Administration Agent digitises credit document review that previously required manual entry, completing it up to 30% faster and with greater consistency. Both are powered by Scotia Intelligence, the bank's unified approach to data and AI, and sit inside systems the commercial team already uses, under the same safety, security and governance standards.
What a month back is for
The headline ambition is bold. By the end of 2027 the bank wants to give every employee back the equivalent of one month of administrative and processing work per year, to spend with clients and on more efficient lending. Chris Manning, Executive Vice President, Canadian Commercial Banking, framed it in relationship terms. "Banking relationships are built on trust and a deep understanding of our clients, who expect a partner that sees the bigger picture and is as invested in their success as they are," he said. "Commercial Navigator gives our teams a stronger starting point for every client conversation and helps speed up turnaround times, so they can spend more time bringing their expertise and insights to our clients."
That framing is right, and it is also the hard part. Time saved by automation has a habit of evaporating. Without a deliberate plan, a freed month becomes more deals per banker, tighter turnaround targets and a busier pipeline. A dashboard might call that productivity, yet it would fall short of the deeper client relationship the bank is describing. A month back is a promise about what bankers will do with that month, and promises need someone to keep them.
From preparing analysis to interrogating it
The more interesting shift is in the work itself. When an agent produces ratios, trends and suggested questions in under a minute, the banker's value moves from assembling the analysis to challenging it. Does the trend reflect the business or a one off? Is the suggested question the one that matters for this client right now? "Commercial Navigator brings our people and AI together to deliver better outcomes for our clients," said Luke Gee, Executive Vice President and Chief Data and AI Officer. "By taking on routine administrative work, it enables our teams to focus their judgment where they add the most value, while helping clients get answers faster."
Judgement is the operative word. A banker who reads the agent's questions aloud to a client has added little a template could not. A banker who treats them as a starting point, then follows the answer somewhere unexpected, is doing the work the bank says it wants more of. There is a quieter tension too. The manual hour was where junior bankers learned what a healthy set of accounts felt like, and that intuition will need another way to form.
What this means for L&D
For learning and sales leaders, the task is to plan for the time dividend before it arrives. Two capabilities matter most. The first is judgement on agent output: knowing when to trust a ratio, when to question it and how to spot what the model has missed. The second is client conversation skill, the ability to turn a list of prompts into a discussion that surfaces what a business owner is actually worried about. Neither develops simply because a tool was deployed, which is why organisations rolling out similar agents should pair them with structured practice of the kind we build into our AI Training. Then measure the right things, including whether conversations got better and not only how fast credit moved. If the month disappears into volume, the agents will have worked and the strategy will not.
Sources:
News and insights for innovation, digital transformation, future of work and L&D leaders.
Stay ahead of learning and development, corporate innovation and digital transformation news. Plus the future of work. For leaders in AU, NZ, HK, SG, the US, the UK and Canada.





