Atos launches Polaris AI Platform

French HQ’ed system integrator Atos has become the latest technology supplier to launch its own AI Platform, as it seeks to take advantage of the demand for the development, deployment and management of enterprise-grade autonomous AI agents.

The Atos Polaris AI Platform includes six pre-built AI agents delivering efficiency gains across enterprise functions, with Atos outlining potential results for its different agent types:

• Its AI Developer agent reduces software development efforts by 40-50%, autonomously analysing business requirements and orchestrating solution development.
• The Quality Assurance agent enables end-to-end testing orchestration, validating requirements, generating test cases, and creating reports independently, achieving 50-60% reductions in efforts and lead times.
• The IT Support Engineer agent provides automated ticket analysis and resolution, conducting in-depth log file analysis to determine root causes and recommend solutions, delivering 25-35% effort reductions.
• The Contract Analyst agent continuously monitors contracts for compliance risks, flagging potential breaches and recommending corrections, reducing review cycle time by 30-40%.
• A Financial Reports Analyst agent interprets large financial documents to provide accurate summaries and actionable recommendations whilst cross-validating for anomalies, delivering 50-60% productivity improvements.
• Its Market Researcher agent leverages trusted organisational data to perform in-depth analysis, synthesising findings in business-tailored formats and enabling 60-70% reductions in research efforts.

There are not any details on the underlying AI models used to power these agents, but we can expect the usual suspects. The agents being offered provide a good spread of capabilities, broadly in line with what we see elsewhere, though as ever we have to take any stated productivity gains with a degree of scepticism as results always vary by organisation. Some further investigation will also be required to understand just how autonomous the AI agents are, and where they hand off to human actions, as we have seen a long of ‘agentic’ washing across the market.

This seems like a very positive move from Atos though, and what really caught my eye was the incorporation of "Agent Ops" functionalities, which ensure AI agents align with business key performance indicators through compliance, performance, and cost management practices. It also lists multi-agent collaboration and coordination through standard Agent-to-Agent protocol, a key feature as organisations grapple with managing and integrating multiple agents across different business lines.

Posted by: Simon Baxter at 09:40