Security company Tanium on Aug. 27 unveiled new features in its next-generation agentic AI system, Tanium Atlas, at its annual conference, Converge Seoul.

Tanium Atlas is a native agentic AI system built into the Tanium platform. It analyses risks in IT and security environments based on real-time endpoint data, prioritises them, and then, after operator approval, links prioritised tasks to those items.

The company said operators can build a single-workflow environment that starts with questions in natural language and leads to problem resolution, instead of analysing complex tools and query syntax or separate data sources. Tanium Atlas also offers a Sensor Creation function that lets operators create and validate custom sensors in minutes to extract tailored data from endpoints without specialised knowledge. It supports other functions including Template Gallery to share reusable validated pages, Reports & Exports to export panel, page and audit data in PDF, CSV, email and HTML formats, and Activity Center to view key recommendations, alerts and approval items on one screen.

It also includes a Background Agents function that performs long-running tasks even when operators are not logged into the console. The company said Memory personalises Atlas for each user to provide tailored experiences and speed up responses based on learned past patterns, and that Atlas can also integrate with various AI clients through an Atlas MCP Server.

Steven Yang (스티븐 양), senior director of product management and portfolio leader at Tanium, said, "Atlas adds intelligence to fast, real-time endpoint data that only Tanium can collect at scale, understands operator intent and responds through accurate parallel execution."