Cognite’s latest product release advances our mission to deliver $100B in realized customer value by 2035. Scaling industrial workflows that optimize production, reduce unplanned downtime, and increase worker efficiency requires unified industrial data in context, AI agents that can take trusted and autonomous actions, and industrial experiences that reimagine how agents and people can deliver value together. While every Cognite release reinforces the foundational trust, security, and scalability of the Cognite Industrial AI and Data Platform, this release further accelerates industrial workflow transformation.
To illustrate the impact of workflow transformation, one industrial organization working with Cognite used to manually review unified operator rounds, maintenance logs, shift notes, and alarms before each shift, a process that could take hours if something was amiss. By unifying this information in context, an AI agent can now identify and prioritize areas of concern in minutes, reducing unplanned downtime by elevating critical issues before they impact production.
By anchoring AI agents in a trusted data foundation, industrial organizations are moving beyond dashboards and manual analysis to true self-service industrial intelligence where insights, actions, and decisions happen in context and in real time.
Industrial experiences: Delivering Operational Impact with AI Agents
With this release, we are further enhancing the industrial experiences to make insights more intuitive when working alongside AI agents. Subject Matter Experts (SMEs) deserve to analyze information as intended, enabling faster troubleshooting and more efficient execution.
Correlating data trends with real-world events has historically been a time-consuming, error-prone challenge requiring SMEs to cross-reference multiple screens, logs, and systems. With this release, users can now overlay events, such as work orders, directly onto time series data in Cognite Charts, providing a single, contextualized view that eliminates manual correlation and ensures both users and agents operate on the same source of truth during analysis.
With the necessary operational context in place, users can work with an agent to perform calculations, identify deviations, isolate root causes, and drive deeper analysis. By automatically linking maintenance records with live sensor data through the industrial knowledge graph, Cognite keeps critical operational information continuously synchronized, preventing data gaps and the drift that occur when systems are not connected in real time.
In addition, this release continues to elevate the 3D experience. When using traditional solutions, navigating multi-level 3D environments can feel like being dropped into a maze with little or no context.
Now, an engineer can see a 360-degree view of an asset’s surroundings the moment they search for it, confirming they have the right location before opening the full model. This reduces time spent orienting in complex environments and minimizes the risk of acting on the wrong asset. To further streamline navigation, the 3D MiniMap shows the current camera position and allows users to jump to different locations with a single click without long wait times, whether they are in the field or the office.
Atlas AI: Driving automation with AI agents
With this release, Cognite Atlas AI becomes more responsive, secure, and natively integrated into your daily operations. Atlas AI industrial agents are intent-based engines that understand operational intent and user direction. They extend beyond conversational assistants and accelerate decision-making and workflow execution.
Atlas AI Agents now retain session history, allowing users to continue workflows across multiple applications over multiple sessions. With persistent conversations, users can return to previous sessions without needing to reestablish context. For example, planning a maintenance schedule can continue to be revised throughout the week as new information becomes available. Additionally, with conversation history, users can now work across multiple workflows in parallel without losing context.
For Atlas AI agents to operate safely in high-stakes environments, we offer per-agent granular access control. No need to rely on blanket "user-role-level" permissions, where managing security for individual AI agents is an "all-or-nothing" process. Cognite delivers precision governance directly to users. Within the Atlas AI agent workbench, users can assign authorized co-owners to modify an agent’s configuration or core logic. This ensures business continuity, allowing teams to maintain collective agent capacity and remove individual bottlenecks.
Industrial knowledge graph: a trusted foundation for scalable industrial workflows
The Cognite Industrial Knowledge Graph is the trusted foundation that delivers a dynamic and interactive experience for users and agents. It extends beyond simple connections and provides rich context within sources, organizing it in a way that reflects how industrial operations actually work. Process engineers need to know where a pump is located on a P&ID and see the live time series data in one view.
This release improves automatic tag detection via diagram parsing, ensuring diagrams can be quickly contextualized and never treated as static files. The equipment, systems, and relationships documented in the diagrams are recognized and automatically linked to operational data, maintenance history, and related documentation. Each parsed diagram enriches the industrial knowledge graph, adding context and accuracy over time, without manual effort.
All groups do not use the same equipment naming structure, and the new search functions accommodate the different naming conventions used across the site.
- User from Major US Refinery
Further transforming static diagrams into searchable assets with automatic tag detection and intelligent parsing
Cognite Search continues to deliver smarter results through advanced industrial parsing that understands the nuances of equipment naming. For example, by recognizing that "0059", "59", and “00-59” represent the same physical asset, we strengthen user trust through higher relevance for detailed searches, eliminating noisy results, and finding exactly what they are even faster than before.
It’s a night and day difference. The new search is much faster and eliminates the ‘fluff’ of noisy results. Because the engine now parses and separates terms, the user doesn’t need to be precise with the search. What we need is right at the top, and it’s making our operators’ lives easier.
– Brett Green, Operator, Koch Ag & Energy Solutions
With the improved Data Transformation user interface, now every user can visually review their data pipelines. With an expanded metric library, users can clearly see how data moves end to end, pinpoint issues such as data rate limits or resource updates, and take action immediately, without writing the code or waiting for engineering support. This added transparency significantly reduces operational overhead and further empowers teams to resolve issues faster and work more efficiently.
Finally, in this release, we are reducing the efforts to manage data pipeline at scale with automated notifications. Users can subscribe to relevant pipelines and receive notifications regarding the health of a function, transformation, or agent. This ensures users are always operating on the freshest information, while also reducing unnecessary actions and enabling more precise, context-aware decisions.
With the Q1 2026 release, we are accelerating our focus on platform velocity and our journey to unlock $100 billion in customer value by 2035. By grounding every AI action and user experience in the Industrial Knowledge Graph, we are enabling industrial workflows to run at unprecedented speed and with context across every site, helping customers achieve tangible economic results, including a 465% ROI and a payback period of less than 6 months.
Ready to see what this means for your operations? Book a meeting with our team.