How BMO Automated the Slowest Step in AML Screening

Meet BMO

BMO has been in banking since 1817 and serves more than 10 million customers across North America. Its enterprise risk organization owns anti-money-laundering compliance, where regulators dictate not just the answer, but the method.

The challenge

Before BMO takes on a new counterparty, it needs to know what the public record says about them. That includes convictions, enforcement actions, and concerning patterns in the press.

The search was mandatory, but analysts had to run it manually. They typed names into search, reviewed results, and documented what counted. As volume grew, BMO relied on partner firms to handle the overflow. It worked, but it only scaled by adding people.

BMO knew that adding automation into the process would help their analysts move more efficiently. As part of that automation, the bank needed a reliable retrieval system that used high-quality, securetunable search parameters and returned source-backed output.

Solution and why Tavily

BMO found that having Tavily as the retrieval layer allowed for secure web search and faster risk determination from its analysts. Tavily also fit an architecture pattern its security team had already approved, making it an easy choice.

Now, when a BMO analyst searches an entity name, Tavily returns ranked URLs with relevant passages already extracted. When a deeper review is needed, Tavily can retrieve the full -page content.

From there, BMO takes over. Its own models score possible risk, and its investigators make the final call.

Throughout the integration, the Tavily team and BMO technology worked together to build a solution that allows analysts to move through more reviews, faster and with confidence in the findings. One managing director at BMO said that using Tavily as a data retrieval utility also enabled uniformity of retrieval results across different investigator actions in BMO.

Using Tavily as a data retrieval utility enabled the uniformity of retrieval results across different investigator actions at BMO.

Ashutosh Sinha, Managing Director, ERPM Data and AI Technology at BMO

What's next

AML was the first production use case. 

Search volume through Tavily grew from roughly 40 calls in the first month to more than 129,000 in July 2026, with more than 97,000 calls in each of the previous six months.

Because the integration is API-only, BMO's security review only had to happen once. New use cases can build on the same approved foundation, and as different teams build and deploy agents, adoption of Tavily is expanding across BMO.

Talk to our team about what an agentic search layer could take off your analysts' plates.