Enterprise Knowledge Search
Cultivating service-oriented relationships is central to BankUnited's core mission. As such, the company wanted to help its employees more effectively surface information from its policy and procedures system, which consists of approximately 400 documents.
Reported impact
BankUnited has rolled out an initiative, "OpportunitiAI," to help identify other generative AI use cases and align its business divisions. "This solution proved the technology for us," says Morales.
Regional bank BankUnited Inc. (BankUnited) strives to deliver accurate, timely answers to its small and medium business (SMB) customers when they reach out with questions. However, employees struggled to quickly find information, which resulted in lengthy call times and inconsistent responses. To solve this problem, BankUnited created a generative artificial intelligence (AI) model using Amazon Bedrock , which helps build and scale generative AI applications. This application can answer policy-related questions almost instantly, helping BankUnited boost efficiency and enhance customer and employee satisfaction. With SAVI, employees can phrase their questions naturally while getting reliable responses.
"It also reduces the training life cycle of the employees." Empowering its employees has had downstream effects.
"It also reduces the training life cycle of the employees." Empowering its employees has had downstream effects.
"The seamless integration of AWS services streamlined our development process," says Jhon Pereda, AVP BI and analytics manager at BankUnited.
So, BankUnited turned to Amazon Web Services (AWS) to build an AI application: SAVI.
The bank also adopted Amazon Kendra , an intelligent enterprise search service, to help its employees quickly find information.
To solve this problem, BankUnited created a generative artificial intelligence (AI) model using Amazon Bedrock , which helps build and scale generative AI applications.
Using Anthropic's Claude 2 model in Amazon Bedrock, BankUnited achieved an accuracy rate of 95 percent and response times of less than 10 seconds.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
AWS Sample Nov 21, 2024 Discover Financial Services In its various lines of business, including decision and credit and portfolio risk management, Discover faced the challenge of making sure that it could launch its services fast enough. “We wanted to make decisions… The high compute power available was used for parallel model training, which reduced the processing time of 30… Amazon EC2 P3 Instances · Amazon EC2 P4 Instances · Amazon EFS AWS Sample Nov 21, 2024 Sun Life After launching more than 40 generative AI experiments, Sun Life needed to move from experimentation to production value with a secure way for employees to retrieve internal information quickly inside the company’s own… over 10,000 queries a week Amazon Bedrock AWS Sample Sep 1, 2024 SBI Life Insurance SBI Life Insurance, an SBI Insurance Group company, provides consumers with products such as an affordable mortality insurance plan, which allows customers to revise the sum insured based on their stage of life; a… It also shortened the training periods by about 30 percent. Amazon Bedrock · Amazon Kendra · Amazon OpenSearch Service 