{"id":7134,"date":"2025-07-29T11:40:42","date_gmt":"2025-07-29T11:40:42","guid":{"rendered":"https:\/\/staging.elixirrdigital.com\/us\/case-studies\/a-large-scale-us-regional-bank\/"},"modified":"2026-03-16T10:59:45","modified_gmt":"2026-03-16T10:59:45","slug":"a-large-scale-us-regional-bank","status":"publish","type":"case_study","link":"https:\/\/staging.elixirrdigital.com\/us\/case-studies\/a-large-scale-us-regional-bank\/","title":{"rendered":"Reducing time to decision by 88% with a GenAI-powered EDD solution"},"content":{"rendered":"\n
\n
\n
\n
\n Back to Case Studies<\/a> <\/div>\n

\n Case Study<\/p>

\n A Large-Scale US Regional Bank<\/h1> <\/div>\n
\n
\n

Creating a GenAI-powered solution to reimagine the enhanced due diligence process.<\/p>\n <\/div>\n <\/div>\n<\/div>\n

\n \"\" <\/div>\n <\/div>\n <\/section>\n\n\n \n
\n
\n
\n

\n About the brand<\/h2> <\/div>\n
\n
\n

Our client is a large, highly regulated organization responsible for conducting detailed customer risk assessments. With growing data volumes and a need to improve efficiency without sacrificing accuracy, the business was exploring automation to modernize its approach.<\/p>\n

The challenge<\/h2>\n

Their Enhanced Due Diligence (EDD) process relied on manual review of complex customer information to identify risks and produce justifications for internal decisions. This time-intensive process involved gathering data across a fragmented data ecosystem in order to provide expert level human analysis on the level of risk.<\/p>\n

The client wanted to streamline reviews while maintaining control over sensitive and obfuscated data to ensure quality, consistency and compliance.<\/p>\n

The solution<\/h2>\n

We started by mapping out the full EDD process with the client to spot where time and effort were being lost.<\/p>\n

Next, we built a GenAI-powered tool to help automate key tasks (like analyzing customer data and writing up risk reports) so teams could spend less time on manual work and more time applying their expert judgement.<\/p>\n

We used obfuscated sample data to safely train the model without exposing real customer information, making sure it could handle sensitive inputs while staying compliant.<\/p>\n

We tested the tool thoroughly and set it up in a way that makes it easy to use across other similar processes in future.<\/p>\n

The results<\/h2>\n