The bank’s AI Agent Assist, announced Friday (Nov. 6), is designed for in-house use by employees and uses agentic AI to help those workers spend less time navigating systems as they assist clients who contact BMO’s Virtual Connect contact centre about their credit card transactions.
“Each year, BMO’s Virtual Connect contact centre receives more than 500,000 calls from Canadian clients about credit card transactions, with each call currently taking an average of 12-14 minutes to resolve,” the Toronto-based bank said in a news release.
With AI Agent Assist, BMO said it hopes to lower the time spent to resolve client inquiries by 20% to 30% and to process about 40% of eligible credit refund requests automatically.
“As a result, in the case of qualifying inquiries clients could receive their credit as soon as the same day,” the release said.
According to the release, the tool brings “AI-enabled transaction matching, policy retrieval and case preparation” straight to workers for review, helping them respond faster and reducing the need to place customers on hold.
“When clients call us about a credit card transaction, they want clarity and a quick resolution,” said Holly Jackson, who heads Virtual Connect. “AI Agent Assist gives our employees the information and guidance they need, so they can spend less time moving between systems and more time helping the client. For qualifying requests, that could mean receiving a credit as soon as the same day.”
The release noted that BMO designed the new tool with a “human-in-the-loop” principle, to keep people at the center of key decisions.
That mindset is reflected in recent research from PYMNTS Intelligence, which shows that while finance departments have moved swiftly to put generative AI to work, they’ve been much more cautious when it comes to letting the technology act on its own.
Generative AI use for cash flow management surged from 16% in 2025 to 80% in 2026, according to the PYMNTS Intelligence/JPMorgan collaboration “Time to Cash: What Two Years of Data Say About Faster Cash and Rising Risk.” Agentic AI went the opposite direction, dipping slightly from 6% to 5%.
“The gap offers a measure of where finance automation stands,” PYMNTS wrote last month. “AI is becoming common in the systems that analyze cash and recommend what to do next, while autonomous execution remains rare.”
