MongoDB Unveils Atlas Agent Engine to Bring AI Agents Into Production Without a New Stack
TECHWORLD ·
✦ AI Summary
MongoDB announced the launch of Atlas Agent Engine, an integrated execution, memory, and governance layer for production AI agents.
Atlas Agent Runtime and Atlas Agent Memory are available immediately in public preview, with usage-based pricing and support under existing Atlas commitments.
The product is designed so customers can keep their existing models and frameworks while adopting only the memory and governance layer, or the runtime as well.
MongoDB announced the launch of Atlas Agent Engine. Atlas Agent Engine is an integrated execution, memory, and governance layer for production AI agents.
The company said the product supports teams in deploying modular agents into production. MongoDB has introduced a new product that brings together the execution, memory, and governance needed to operate production AI agents.
The product is designed so customers can keep their existing models and frameworks. Customers can adopt only the memory and governance layer on its own, or they can also adopt the runtime.
The Retrieval feature runs on MongoDB Voyag AI. The embedding and reranking models recorded top performance on RTEB. RTEB is an evaluation metric used to reflect real enterprise search performance rather than academic datasets.
Atlas Agent Runtime and Atlas Agent Memory are available immediately in public preview. Usage-based pricing applies to Atlas Agent Runtime and Atlas Agent Memory. This usage can be applied under existing Atlas commitments, so no separate new contract is required. As a result, companies can expand and use their existing infrastructure as is.
Amar Akshat, SVP and head of architecture at Paysafe, said analysts often need to manually gather data from multiple systems when investigating anomalies in payment networks. He said these investigations often come under time pressure. He added that intelligent agents based on MongoDB Atlas Agent Engine are expected to help shorten the time from issue detection to team response, while giving analysts more time to focus on key decisions.
James Governor, cofounder of RedMonk, identified context as the key factor in successful agent use in application development. He said companies are struggling to evaluate, integrate, and manage information from multiple systems in order to implement autonomous agentic work. He added that MongoDB Atlas Agent Engine is designed to embed governance into the development of agentic apps through a single platform for memory and identity management.
Atlas Agent Engine is built on a partner ecosystem that enterprises have long trusted. In this structure, Frontier Labs partners can deliver the optimal model directly where enterprise data resides, while SI partners provide industry expertise and build experience for agent deployment.
Pablo Stern-Plaza, MongoDB's CPO of AI and new products, said companies deploying to production have been forced to choose between adopting a specific vendor runtime and becoming locked into a specific model or cloud, or assembling frameworks themselves while managing governance and memory in-house. He said the launch of Atlas Agent Engine eliminates that trade-off and enables enterprises to secure real-time context for agents, along with security and governance built in from the start. He added that companies can freely choose and run any model, framework, or cloud, and that the goal is to build solutions that work fully with whatever choices customers make, rather than assuming what they will need in the future.
Source: TECHWORLD · Lee Gwang-jae
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407605
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.