MongoDB has launched Atlas Agent Engine, an execution, memory, and governance platform aimed at moving AI agents beyond sandboxed proof-of-concept testing.

While autonomous agents reliably pass preliminary evaluation in sandboxed environments, release engineering teams face major obstacles when moving them into live systems. Reliable deployments require precise data retrieval, persistent state retention, and verifiable administrative controls. Without a unified system, software teams must assemble fragmented utilities that break whenever foundational models or framework libraries change.

Pablo Stern-Plaza, Chief Product Officer for AI and Emerging Products at MongoDB, said: “Organisations that want to put agents in production are being forced into a false tradeoff: either adopt one vendor’s runtime and accept being locked into a model and cloud, or piece together a framework and manage governance and memory on their own.

“With the launch of Atlas Agent Engine, that false tradeoff ends today. Enterprises get the real-time context their agents need, with governance and security built in from the start, and the freedom to run any model, any framework, and on any cloud. We didn’t want to ask customers to predict the future. We wanted to build something that works no matter what they choose.”

Voyage AI embeddings lead RTEB retrieval benchmarks

In typical architectures, agents clear their working memory after each run, forcing quality engineers to assemble custom data pipelines for every test cycle. Atlas Agent Engine builds memory retention directly into the database tier, linking native MongoDB queries to Voyage AI embeddings and reranking algorithms.

On the Retrieval for Enterprise Benchmark (RTEB), Voyage AI models rank among the leading performers by testing accuracy against realistic commercial workloads rather than academic text corpora. This tighter data coupling sharpens contextual recall while reducing token consumption across repeated runs.

Engineering groups can introduce the memory and governance modules on their own or run them alongside the Atlas Agent Runtime. Commercial billing operates on consumption and draws directly against an organisation’s existing Atlas financial commitments, removing the need for separate procurement contracts.

Amar Akshat, SVP of Architecture at Paysafe, commented: “Investigating unusual activity in our payment network today means our analysts stitching together data from multiple systems by hand, often under time pressure.

“We’re excited about the potential for an intelligent agent – built on MongoDB’s Atlas Agent Engine – to shrink the time between a problem emerging and our team acting on it, giving our analysts more time to focus on the judgment calls that matter most.”

James Governor, co-founder of RedMonk, added: “Context is the critical success factor in successfully using agents for application development. Enterprises are currently struggling to assess, integrate, and manage information across multiple systems to enable an ontology for autonomous agentic work.

“MongoDB Atlas Agent Engine is designed to bake governance into agentic app development with a single platform for memory and identity.”

Open standards guide cross-cloud governance and audit

Most production stacks scatter access policies, identity credentials, behavioural guardrails, and budget caps across disconnected consoles. Atlas Agent Engine directs those administrative controls through a single plane. 

Every operation ties to an authenticated identity, whether a human operator or an autonomous agent, governed by runtime rules that administrators cannot switch off. Security verifications that once took weeks to assemble now resolve in seconds. Because memory, data access, and governance run within one operational system, deployment pipelines encounter fewer structural points of failure.

Neutral standards prevent vendor lock-in across continuous delivery environments. The engine supports open protocols including Model Context Protocol (MCP) and Agent2Agent (A2A). Test engineers can swap foundational models or framework libraries through simple configuration adjustments rather than rewriting underlying test scripts.

Workloads execute consistently across hyperscale cloud providers, on-premises systems, or local developer machines. To promote these interoperable architectures, MongoDB joined the Open Secure AI Alliance and the Agentic AI Foundation within the Linux Foundation.

Systems integration partners, including Accenture, are incorporating the engine into their client delivery practices.

Ram Ramalingam, Global Lead of SW Engineering and Head of RDE at Accenture, said: “Atlas Agent Engine brings the enterprise-ready capabilities, context, and constraints needed to help AI agents deliver real-world impact.

“Combined with Accenture’s governance, architecture, and deep industry expertise, it creates a powerful foundation for accelerating AI transformation and delivering outcomes at scale. Our shared commitment to delivery and customer success makes this partnership particularly strong.”

Atlas Agent Engine is available in public preview alongside MongoDB 9.0 and Atlas Infinite.

See also: Artificial Analysis Cyber Index tests defensive AI security models

Want to learn more about cybersecurity from industry leaders? Check out Cyber Security & Cloud Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the AI & Big Data Expo. Click here for more information.

Developer is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.