Cloud database provider MongoDB Inc. today announced a slew of updates to its platform that are designed to help software developers build applications that can harness the power of generative artificial intelligence.

Among other improvements, it announced new Vector Search functionality that improves the way it filters and aggregates unstructured data, making it easier for generative AI applications to access. Besides enabling generative AI, MongoDB is also using it to power more intelligence experiences for developers that aim to improve their productivity when using the MongoDB Atlas cloud platform.

The company also announced a new version of its flagship database, MongoDB Atlas for the Edge. It said that will make it easier to deploy applications that can process data closer to where it’s created.

The new capabilities in MongoDB Atlas, announced at the company’s MongoDB.local London developer conference, should add further momentum to a generative AI push that began earlier this year. Back in June, MongoDB announced a new Vector Search feature, and today it’s building on that foundation.

MongoDB is the creator of the document-oriented MongoDB database, which is used for a wide range of data-intensive applications and prized for its ability to store information in multiple different formats. MongoDB Atlas is the cloud-hosted version of that database.

The company believes the choice of database is critical for enterprises that want to build generative AI applications. It argues that AI developers need to build on a database that’s unified, fully managed, flexible and scalable.

MongoDB says it provides all of these things, and more besides, the most important of which is its enhanced Vector Search platform. Vector search is vital for AI as it provides a way to create numerical representations – vectors – of unstructured data such as images, written notes and audio. It makes this data accessible to generative AI models so they can be trained on it.

MongoDB presents Atlas as a unified solution to process data for generative AI apps, marrying structured data with vectors in a single platform. It launched Vector Search in preview in June, and today it announced significant performance improvements that reduce the time it takes to create data indexes by 85%. In addition, MongoDB Atlas Vector Search now integrates with fully managed data streams from Confluent Cloud, making it easier to access real time data for AI.

One of the new capabilities today enables developers to create a dedicated data aggregation stage with Vector Search using MongoDB’s Query API syntax and Aggregation Pipeline too. This makes it easier to filter results and improve the accuracy of information retrieval, thereby reducing AI “hallucinations,” or the tendency of AI models to generate inaccurate responses.

The indexing acceleration is a key improvement too, because generating vectors is the first step in the preparation of AI training data. Once the vectors are created, developers must build an index for that data to be queried for information retrieval. It’s also necessary to update the index every time data is changed or new data is added.

As for the real-time data streams using Confluent Cloud, it can be helpful in building more engaging and responsive generative AI applications that can learn and respond in real time, MongoDB said. With today’s update, developers can integrate Confluent Cloud data streams within MongoDB Atlas Vector Search to provide their apps with “ground-truth data,” or the most accurate and up-to-date information, from a variety of sources. This will make generative AI apps more responsive to changing conditions, the company said.

Constellation Research Inc. analyst Doug Henschen said the vector search capabilities are focused on the needs of developers and help MongoDB set itself apart from other platforms and grow its adoption. He said the most impressive thing about MongoDB’s vector search updates is that it’s not just announcing new capabilities, but also illustrating how it aids in innovation.

“MongoDB is sharing the real-world customer examples of Dataworkz, Drivly, ExTrac, Inovaare Corp., NWO.ai, One AI and VISO Trust along with plenty of details on how it’s helping them to innovate,” Henschen said. “This makes the announcement come across as much more real, and will inspire other customers to begin experimenting.”

New generative AI experiences for developers

While MongoDB is making it easier for developers to build more sophisticated generative AI apps, it has also built its own generative AI tools for them to use. The new developer experiences announced today make use of generative AI to improve productivity, reducing the time and effort they spend on mundane tasks so they can focus on the more taxing aspects of their work.

The company announced a new experience for MongoDB Compass, which is now able to generate queries and aggregations from natural language to build data-driven applications faster and more easily. Meanwhile, MongoDB Atlas Charts users can now create rich data visualizations for business intelligence using natural language prompts.

MongoDB Documentation gains a new generative AI chatbot that provides rapid answers to technical questions, while MongoDB Relational Migrator can now convert Structured Query Language to MongoDB Query API Syntax to help automate database migrations.

Henschen said MongoDB’s new generative AI-powered features are mostly in line with what other database platforms are doing, using the technology to improve developer productivity. “All but the chat feature are currently only in preview, but it’s still a good set of capabilities that shows MongoDB is not conservative when it comes to infusing its products and services with the latest generative AI tools,” he added.

MongoDB Atlas lands at the edge

With MongoDB Atlas at the Edge, the company is bringing its advanced database functionality to edge environments, making it simpler for organizations to deploy their apps closer to where data is generated.

It will enable companies to deploy distributed applications that can reach end users faster, providing real time experiences, the company said. MongoDB Atlas for the Edge makes it possible for data to be stored and synchronized in real-time across data sources and locations, and will provide advantages in use cases such as connected cars, smart factories and supply chain optimization, the company said.

“MongoDB has long had its mobile database and synchronization capabilities, but with Atlas at the Edge it’s clear that it’s now driving more consistency and continuity, helping to simplify edge-to-cloud use cases with two-way interactivity,” Henschen said.

Image: MongoDB