Sibos 2025: AI agents help banks unlock the power of data, say panellists at Sibos 2025

Sibos 2025: AI agents help banks unlock the power of data, say panellists at Sibos 2025

Microsoft

From left: Microsoft’s Kathleen Woodward, Scotiabank’s Matt Loos, EY’s Diana Halder and Microsoft’s Santhana Sankaramurthy

Experts from Microsoft and EY explained how they worked with Scotiabank to automate reconciliation reports using AI agents, cutting hours of manual processing and giving employees more time to focus on higher-value tasks

Alice Chambers

By Alice Chambers |


AI is helping banks unlock the potential of their data, streamline operations and deliver more meaningful client insights, according to industry leaders speaking at Sibos 2025.

The session brought together Kathleen Woodward, head of banking, industry advisory at Microsoft Americas; Matt Loos, managing director of global payments at Scotiabank; Diana Halder, partner for North American payments at EY; and Santhana Sankaramurthy, director of AI business solutions at Microsoft. The panellists discussed how banks can take a pragmatic approach to AI adoption and shared lessons from a joint Scotiabank, EY and Microsoft project.

Loos outlined Scotiabank’s challenge: producing client reconciliation reports could take up to 20 hours, as the raw data was not broken down into adjustable fields. “Reports are painful and time-consuming,” he said. “We struggled with how to tell clients how to repair issues. Now we have an agent that cleans the data, structures it so systems can talk to each other, and matches data to create client insights reports. That allows us to scale something that was extremely manual.”

He added that by applying a small set of rules and feeding data into a knowledge centre, the AI agents quickly became more intelligent. “We proofed it out in two weeks rather than a year and a half,” Loos said. “It was low risk for us – all post-transaction, not touching a live action – the perfect use case. Once you learn how agents can think on their own, the more you can do.”

Sankaramurthy added: “The challenges we encountered with Scotiabank were mainly about heavy lifting with data so the rest of the workflow worked out. If we hadn’t got it right, we couldn’t have fixed it very easily. Now that we have the solution, we’re looking to expand horizontally and vertically – getting the product into production.”

Halder said the collaboration shows how banks can apply AI beyond customer-facing chatbots, which are already common across the industry. “Helping Scotiabank was about looking inside – adding value to the end user and upscaling the operations team,” she said. “The bank deployed AI in the back office, and that really makes it frontier.”

She also shared advice for others considering similar projects: engage stakeholders early, start with a straightforward application and focus on realisable value. “Be realistic – cost reduction needs to be involved,” she said.

The panellists agreed that AI agents open up new opportunities for scaling innovation across financial services. Loos suggested banks should prioritise simple use cases. “Find a use case – the simpler the better. Money is number one, number two is what do we have the capacity to build.”

Employees at Scotiabank have already felt the impact, he added. “Those who would’ve done the reports provided feedback. It’s interesting to see the power that these agents put back in their hands.”

Woodward highlighted Microsoft’s own efforts, pointing to Finance in Microsoft 365 Copilot, which uses agents to autonomously generate reports. Built in just four weeks, with two weeks of testing, it will be available soon. She also noted that Microsoft used its reconciliation agent internally, acting as “customer zero.”

Contact author

x

Subscribe to the Technology Record newsletter


  • ©2025 Tudor Rose. All Rights Reserved. Technology Record is published by Tudor Rose with the support and guidance of Microsoft.