MongoDB has announced the public availability of its Atlas Vector Search integration with Amazon Bedrock. This collaboration allows developers to sync their foundation models and AI agents with the proprietary data stored within MongoDB, resulting in more accurate and personalized responses through the use of Retrieval Augmented Generation (RAG).
Sahir Azam, MongoDB’s chief product officer, highlighted the importance of ensuring accuracy and protecting proprietary data when using AI-powered systems. He stated, “We’re making it easier for joint MongoDB-[Amazon Web Services] customers to use a variety of foundation models hosted in their AWS environments to build generative AI applications that can securely use their proprietary data within MongoDB Atlas to improve accuracy and provide enhanced end-user experiences.”
Amazon Bedrock is AWS’s managed service for gen AI, offering enterprise customers a central repository for all their AI app-building needs. While there is a growing collection of models available from external parties, companies may prefer to leverage their own databases to gain greater context about their customers.
This is where MongoDB’s integration becomes valuable. Developers can customize the foundation model of their choice with their own data privately. Scott Sanchez, MongoDB’s vice president of product marketing and strategy, emphasized the importance of using real-time operational data in the models to avoid generic responses. He explained, “You can build these gen AI applications, but unless you can put your own real-time operational data into the models, you’re going to get generic responses.”
The integration between MongoDB and AWS simplifies the process of connecting the dots and allows customers to privately customize their large language models with proprietary data. For example, a retailer could develop a gen AI application that uses autonomous agents to handle tasks like processing real-time inventory requests or managing customer returns.
This collaboration between MongoDB and AWS is not their first. MongoDB’s Vector Search is already available on Amazon SageMaker, and Atlas is supported by CodeWhisperer. MongoDB has also introduced its AI Applications Program (MAAP) to assist enterprise customers in creating AI applications.
In summary, the integration of MongoDB’s Atlas Vector Search with Amazon Bedrock enables developers to use their own data to create customized gen AI applications. This collaboration addresses the concerns of accuracy and data protection while providing enhanced end-user experiences. By leveraging their own databases, companies can gain valuable insights about their customers and avoid generic responses. MongoDB and AWS have been working together to make AI application development more accessible and efficient for enterprise customers.