Atos unveiled at Microsoft Ignite 2025 a clear move toward advanced data team automation: the launch of Autonomous Data & AI Engineer, an agentic AI solution built on Microsoft Azure that aims to take on much of the heavy lifting in data engineering and AI.
The offering, supported by the Atos Polaris AI Platform and Microsoft’s cloud and AI services, aspires to transform many tasks currently handled by data engineering, BI, and MLOps teams into orchestrated workflows managed by AI agents capable of planning, executing, and supervising complex processes with minimal human intervention.
A “squad” of AI agents to accelerate data projects
This new solution is presented as an “autonomous data and AI engineer” composed of multiple specialized agents working collaboratively. Its role is to handle end-to-end typical data engineering tasks:
- Ingest structured and unstructured data from various platforms.
- Apply quality rules and transformations.
- Build views and data models ready for analysis and consumption.
- Expose these data to other AI agents and visualization tools to generate business insights.
Practically, this means that a significant portion of what data teams today do—such as loading files, designing pipelines, creating intermediate tables, documenting changes—can be delegated to a set of agents that understand the context, execute sequential steps, and return results in a reviewable format.
Atos claims that this approach reduces manual effort and speeds up the deployment of data operations by up to 60%, in addition to cutting operational costs by around 35% thanks to DataOps agents that decrease the average incident resolution time.
Direct integration with Azure Databricks and Snowflake on Azure
In its initial version, Autonomous Data & AI Engineer is available for two of the most widely used data platforms in the Microsoft ecosystem:
- Azure Databricks, focused on advanced analytics and lakehouse architecture.
- Snowflake on Azure, the popular data cloud deployed on Microsoft infrastructure.
The agents can connect to these environments, load data from external sources, apply transformations, and generate business views without requiring users to write SQL code or complex notebooks. Once this data “skeleton” is built, analysts and business users can rely on other AI and visualization agents to ask questions in natural language, create dashboards, or validate hypotheses interactively.
This approach aims to ease the burden on core data teams, which often become bottlenecks in large organizations by concentrating the technical expertise needed to deploy new use cases.
Atos Polaris AI Agent Studio: orchestrate agents without coding
The core of the offering is the Atos Polaris AI Agent Studio, a no-code environment integrated within the Polaris platform that enables technical and business profiles to:
- Compose and orchestrate multiple AI agents within a single workflow.
- Connect them with language models (LLMs), external tools, and other agents.
- Use open standards such as Model Context Protocol (MCP) and Agent-to-Agent (A2A) protocols, designed to enable different agents and applications to collaborate seamlessly.
For example, a business user could drag and drop components representing:
- An agent ingesting data from a CRM.
- Another cleaning and enriching the information.
- A third constructing a purchase propensity model.
- An visualization agent presenting results on a dashboard.
All without direct programming, yet allowing data teams to add their own rules and validations.
Atos describes this approach as a step toward the “Services-as-Software” paradigm: instead of contracting consultancy hours for repetitive tasks, companies consume packaged solutions in the form of autonomous agents that can be deployed, configured, and monitored from their own platforms.
Agentic AI aligned with Microsoft’s responsible AI principles
An element that Atos emphasizes is that its autonomous data and AI engineer is based on Microsoft’s responsible AI principles. The solution is integrated with Azure services and aligns with Microsoft’s governance, security, and compliance frameworks for enterprise AI use.
This has practical implications:
- Traceability of the actions performed by agents.
- Options to establish data access and usage policies through Azure security tools.
- Mechanisms for reviewing and auditing automated decisions, which is especially important in regulated sectors like banking, healthcare, or government.
In the context of the rise of agentic AI—models capable of decision-making, tool invocation, and coordination—such guarantees are becoming essential for large organizations to confidently deploy autonomous solutions in critical processes.
Fewer tickets, more innovation: what the company gains
Atos positions Autonomous Data & AI Engineer as a direct response to a specific issue: many data departments are overwhelmed by repetitive tasks, integration tickets, and report requests, leaving little room for high-impact innovation projects.
The company states that deploying specialized agents on the Polaris AI platform and Azure enables:
- Reducing dependence on central teams: business units can activate new data flows without waiting weeks for IT capacity.
- Accelerating time-to-market: automating much of the technical work speeds up moving from idea to prototype and from prototype to production.
- Lower operational costs: reductions of up to 35% in data and support operations costs, thanks to automated DataOps.
- Freeing up R&D capacity: time previously spent on “firefighting” can now be dedicated to higher-value use cases like advanced predictive models or personalization.
All this without displacing human teams, but repositioning them toward oversight, use case design, data governance, and validation activities.
Another piece in the Atos–Microsoft two-decade partnership
The announcement follows over 20 years of collaboration between Atos and Microsoft in cloud services, data center modernization, and cybersecurity. With Polaris AI, Atos is building a platform of agents that has already expanded into other ecosystems, such as AWS Marketplace, and now integrates deeply with Azure.
At Microsoft Ignite 2025, the company showcased live how these data and AI agents work alongside human teams in what they call “hybrid squads”: combinations of traditional engineers and autonomous agents sharing tasks, each doing what they do best.
For Atos, the message is clear: the next wave of digital transformation will not rely solely on isolated language models, but on deploying full AI agent systems with governance and open standards capable of operating on the large data platforms already in use across enterprises.
Frequently Asked Questions about Atos Autonomous Data & AI Engineer
What exactly is Atos’ “Autonomous Data & AI Engineer” and what problems does it solve?
It’s an agentic AI solution built on the Atos Polaris AI platform and Microsoft Azure, combining several specialized agents capable of end-to-end data and AI engineering tasks. It ingests data from multiple sources, applies quality and transformation rules, builds data views, and enables insights through other AI agents and visualization tools. Its goal is to reduce repetitive manual work, speed up data pipeline deployment, and ease pressure on data engineering teams.
How does the “autonomous data and AI engineer” integrate with Azure Databricks and Snowflake?
Currently, the solution is available for Azure Databricks and Snowflake on Azure. The agents connect to these platforms to load data from external sources, apply transformations, and generate analysis-ready views without requiring users to write code. Subsequently, other AI agents can respond to natural language questions, generate reports, or feed analytical models, all benefiting from Azure’s security, logging, and governance infrastructure.
What is agentic AI, and what roles do MCP and A2A protocols play in the Atos Polaris AI platform?
Agentic AI refers to systems composed of agents that can plan, reason, invoke tools, and collaborate to achieve objectives. In Polaris AI, these agents coordinate using open standards like Model Context Protocol (MCP), which defines how models access context and tools, and Agent-to-Agent (A2A) protocols, which enable different agents to share information and tasks. This facilitates building interoperable ecosystems rather than closed, isolated solutions.
What specific benefits does this solution offer for enterprise data engineering on Azure?
For organizations already using Azure Databricks, Snowflake, or other Azure data platforms, Autonomous Data & AI Engineer provides a way to automate DataOps and data engineering tasks without rewriting existing architectures. It promises up to a 60% reduction in development and deployment time, operational cost savings of up to 35%, and decreased dependence on central data teams. Being aligned with Microsoft’s responsible AI principles also makes it easier to meet security, privacy, and compliance requirements, especially in regulated sectors.
Sources:
- Official press release from Atos: “Atos Announces the Availability of Autonomous Data & AI Engineer, an Agentic AI Solution on Microsoft Azure, Powered by the Atos Polaris AI Platform,” November 18, 2025.
- Atos corporate website and documentation on the Atos Polaris AI Platform and Polaris AI Agent Studio.