United StatesFinance and insurance2025 – ongoingOceanFirst Financial
Copilot for anti-money laundering due diligence and bond portfolio review
OceanFirst cut bond portfolio reviews from 6-8 hours to about 15 minutes with Microsoft Copilot
Problem
OceanFirst, a bank with $13.3 billion in assets, handled 100 to 200 enhanced due diligence cases a day. Working out the true nature of a complex entity could take more than 6 hours. Bond portfolio reviews took 6 to 8 hours.
Approach
About 150 employees were onboarded to Microsoft Copilot. Investigators upload transaction data from Fiserv and external data from Verafin to judge whether activity is high risk. Copilot also compares bond portfolio rates using sources such as Moody's. The bank also uses Databricks with a model context protocol connection so staff can ask large language models questions about deposit and loan data. Databricks for bond analysis was still in beta at the time of the report.
Results
Takeaway
The bank spent about a year and a half cleaning its data before these tools gave useful answers, according to its CIO.
Categories
- Industry
- Finance and insurance
- Business function
- Finance, risk and legalInternal work and knowledge management
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Company's own claim
Sources
Source summaries
American Banker describes how OceanFirst uses Copilot for anti-money laundering due diligence and bond monitoring, and Databricks to query deposit and loan data. It gives the bank's time savings, the number of Copilot users and the daily case load. The bank's CIO said data cleanup took about a year and a half.
Full details are in the original sources.