Cognite's Q2 2026 release marks a pivotal step in our mission to deliver $100B in realized customer value by 2035. Serving as the foundation for this transformation, the rollout of Cognite Flows™ earlier this May delivers on our commitment to heavy industry by closing the critical gap between complex industrial data, AI, and immediate operational action.
Scaling workflows that optimize production, reduce unplanned downtime, and improve worker efficiency requires industrial data in context, AI agents capable of taking trusted action, and collaborative experiences that help humans and AI work together to deliver measurable value.
With more than 60 new features across Cognite’s platform, our Q2 remains focused on customer value, advancing the industrial user experience, agentic AI, and industrial data foundation. From frontline operators to digital teams to enterprise developers, this release empowers teams to turn domain expertise into immediate action while ensuring the data behind every decision is secured, trustworthy, and interoperable.
This is industrial AI moving from specialized capability to enterprise-wide scale.
Cognite Flows: The action layer for industrial workers
Cognite Flows is the action layer of the Cognite Industrial AI and Data Platform, purpose-built to close the gap between industrial data and operational action. Flows is transforming how industrial teams work with traditional industrial software, adapting to the worker instead of forcing the worker to adapt to the software.
There are three ways to engage with Cognite Flows, designed to serve every role, every interest, and every use case.
Custom Applications
With Custom Apps, certified builders can deploy tailored, production-grade applications 100x faster than traditional software development methods and seamlessly scale them across sites, teams, and the enterprise. A low-code development environment streamlines the full build-and-deploy lifecycle, reducing configuration overhead and freeing up more time to build solutions and address more use cases.
"One of the most urgent challenges at our production sites has been how to effectively capture the specialized knowledge and experience accumulated over many years and transform it into a digital legacy. The applications built using Cognite Flows represent a fundamental transformation that goes far beyond simple visualization tools. By leveraging real-time knowledge graphs, we expect these tools to evolve into proactive AI agents capable of managing complex operations."
– Naruki Akiyama, Advanced System Development Group Leader, System Innovation Technology Office, Technology & Engineering Center, Idemitsu.
Watch the Cognite Flows webinar

- Each custom solution deployed within the Cognite Platform undergoes a certification process to ensure it meets Cognite’s standards for security, reliability, data integrity, and end-user value. This helps customers scale applications with confidence, knowing each solution is built to perform safely in complex industrial environments.
- Cognite provides enterprise-grade hosting, including underlying infrastructure and solution lifecycle management, delivering native authentication, built-in governance, observability, security protocols, and long-term support.
- Teams can build Custom Apps with Cognite, with a partner, or on their own. Cognite provides certification training and review to ensure that all Custom apps are performant, secure, and trusted to address complex operational challenges.

Native Applications
With this release, we are further enhancing the Native Applications experiences, including Industrial Canvas, Search, and others, to integrate seamlessly with the broader Platform experience and improve frontline productivity.
Our stamps library in Industrial Canvas has expanded with over 300 standard industrial symbols categorized for quick discovery during markup and highlight workflows. Frontline teams can now seamlessly use these symbols for P&ID markups when creating maintenance work packages or showcasing complex operational processes directly within Canvas — drastically reducing preparation time and breaking down cross-team communication silos. Usability is further enhanced by advanced features such as proportional resizing and persistent color memory. Plus, new alignment with project settings configurations ensures a completely synchronized and uninterrupted user experience when switching between Canvas and Search.

With mobile-first observation workflows in InField, frontline workers have a more adaptive experience on their mobile screens. Delivering immediate context exactly where the work happens, this capability drives faster onsite troubleshooting, accelerates the review of existing field notes, and drastically reduces duplicate reporting.
To support our commitment to no vendor lock-in, we have radically simplified how users extract information from the Cognite Platform for offline work and for use with third-party tools. When users find and filter data in Cognite Search, they can now instantly download those exact results as a CSV file, preserving the custom filters and layout they created. This gives teams the freedom to autonomously access the data they need for spreadsheets, reports, work order lists, asset reviews, and other offline tasks without submitting IT data requests. By making it easier to move trusted industrial data into the tools teams already use, this update reduces bottlenecks and enables faster, more flexible collaboration beyond the platform.

Adaptive Experiences
Available in private preview in Q2, Adaptive Experiences provide a personalized entry point to the Cognite Platform, tailored to each user and their work across the Platform, including Cognite Flows. Powered by Atlas AI, this capability introduces a configurable landing page that lets users select, create, and arrange widgets based on their role, operational priorities, and day-to-day responsibilities.

Users interact with an AI agent to build their own operational overview, helping them get directly to the insights, workflows, and actions they need as soon as they sign in. This includes embedded charts, custom apps, and live data summaries brought together in a single view. Adaptive Experiences are expected to reach production and become available to every Platform user in Q3 2026.
Cognite’s market differentiation is now even stronger as Flows pulls together several layers into a single proposition: contextualized data, AI agents, user experience, custom apps, and a governed rollout.”
– Sebastian Trolli of Frost & Sullivan
Atlas AI: Major updates with industrial agents are now part of how users already work
With this release, we are launching a powerful set of Atlas AI capabilities designed for agent data analysis, packaging, and organizing domain expertise, and providing a persistent user experience across the Cognite Platform.
With the new Omnipresent Agent Experience, Atlas agents are integrated with full context across the Cognite Platform. With the "Ask Atlas AI" sidebar, users can engage with an agent from anywhere — the Home page, Canvas, or Charts — and it stays open as they navigate, carrying conversation context with them. This is an out-of-the-box, pre-configured agent that is available to all Atlas AI customers immediately, with no setup required, so teams can start using Atlas AI without any building or configuration.

The underlying capabilities have also made significant strides. Atlas can now run data analysis against datasets of up to 5,000 instances ( a 50x increase over the previous limit) by writing and executing code directly against query results. Industrial questions rarely have simple answers; being able to ask "rank the highest-priority maintenance orders from the last month" and receive a complete, reliable result expands the scope of what AI-assisted operations can deliver.
To complement this, Atlas Skills introduces a simpler way to manage domain expertise. Rather than loading all context into an agent's configuration, which can impact result quality as complexity grows, Atlas Skills packages expertise into reusable, independently maintained capabilities. Users can enable specific skills per agent based on user intent, keeping configurations lean and outputs sharp, supercharging rapid assembly of AI agents by composing existing skills rather than starting from scratch.
Finally, with the next generation of the Atlas AI query tool, agents can now navigate complex industrial data models more independently and discover the right data faster, without configuration. For users, this means less setup work, faster access to answers, more tasks completed, and more scalable AI workflows. By helping teams move from question to action faster, this update makes it easier to apply AI to the day-to-day work that keeps operations moving.

Industrial knowledge graph: a trusted foundation for scalable workflows and industrial worker experiences
Data is only valuable if teams can trust it. With this release, Cognite further strengthens the Industrial Knowledge Graph by helping transform complex industrial data and relationships into clean, reusable, and business-ready data products tailored to specific operational needs. This gives teams the trusted foundation they need to build scalable workflows, deliver better user experiences, and extend the value of the Cognite Platform across the enterprise.
Cognite’s Industrial Knowledge Graph is not just an abstract network of data; it is an intelligent map of operations that connects digital information to physical reality. For asset-dense environments, such as refineries, chemical plants, or pharmaceutical facilities, spatial data provides the physical context needed to understand where assets are, how they relate to one another, and what is happening around them, even when legacy CAD models are incomplete or unavailable.
With the new 360-degree image experience, teams can connect visual files directly into the data model, turning imagery into interactive spatial context. Capabilities such as 3D cursoring on underlying point clouds and ground-level navigation markers help operators and remote experts navigate complex environments with greater accuracy and less visual obstruction. This makes it easier to conduct virtual inspections, investigate issues with full operational context, reduce unnecessary site walks, and accelerate time-to-resolution during incidents.
The release also includes several platform updates that make the Industrial Knowledge Graph faster, more secure, and easier to maintain.
With new granular access controls, organizations can now give teams permission to run approved automated tasks without granting broader access to modify data pipelines, helping teams run routine operational workflows more efficiently while reducing unnecessary access and enforcing governance requirements.
Also, API-based email alerts for hosted extractors give teams a faster way to detect and respond to extraction issues. Instead of manually checking pipeline statuses, data engineering teams can configure automated email notifications when specific hosted REST extractors enter a failure or error state. The result is faster failure detection, helping teams act before data gaps affect downstream workflows or production processes. Other new capabilities improve simulation throughput and allow configuration updates to take effect instantly without manual restarts. Together, these updates reduce operational overhead and help ensure trusted industrial data stays current, reliable, and ready to power workflows at scale (see the full release note on hub.cognite.com).
The Q2 2026 release is a coordinated push to put industrial AI within reach of every person in the organization, from the operator starting their shift with a personalized application, to the domain expert querying thousands of maintenance records through Atlas AI, to the digital teams scaling successful workflows across sites and the enterprise.
Powered by industrial data in context, this release helps people and AI work together the way they should: with trusted data, clear context, and the ability to move from insight to action. This is how industrial AI moves beyond promise and into production at scale, delivering operational and business impact.
Ready to see how you can scale industrial workflows and enable your frontline? Book a meeting with our team.

