Industrial AI company Cognite staged its annual practitioner, customer and partner conference in Houston, Texas this week. With a focus on automating, integrating and accelerating business operations, processes, mechanics and driving agentic AI, the company welcomed speakers from customers and partners to showcase its work, which it explains as a mission to offer differentiated value through core data modeling and AI-driven innovation.
Cognite chief executive officer Girish Rishi suggests that the world’s industrial landscape is experiencing a shift away from IoT platform-centric approaches to flexible, scalable solutions that integrate seamlessly across operations. How can data science teams follow and flow with this trend? He says that Cognite Data Fusion serves as the “foundational platform” for industrial AI and data solutions.
In a keynote entitled “From Vision to Value – Accelerating AI Transformation”, Rishi explained his vision for the company he leads and the platform that his team is taking to market.
Energy: the ‘anchor-vertical’
“Although our focus is industrial, we work with a focus that is multi-vertical. We will always work with energy as our anchor-vertical and our focus in this market is on performance and mission-criticality as we work to fuel artificial intelligence deployments on upstream, midstream and downstream assets,” said Rishi. “Today, we take the responsibility to drive value for businesses very seriously. This show is all about providing content that welcomes customer speaking opportunities in every single session; they are the best storytellers after all.”
Talking about the time he has spent with customers and partners all around the world from Norway to the Middle East to Japan, Rishi says that the renaissance for industrial AI is a “patriotic call” that sees countries embracing sovereignty as they centralise their concerns onto how AI models are trained from the point of view of a) performance and also b) with a view on security as a pressing concern.
Unlock the workforce, unlock the value chain, unlock the wider ecosystem. This was the summary triumvirate tagline that Rishi’s keynote hinged around… and it served to set up the customer showcase sessions that Cognite Impact 2025 was designed to deliver. Talking about the way users in the field can use the Cognite knowledge graph (working with Cognite Atlas AI agents) to assess the status of real world issues, this event featured live demos of agentic analysis that might be carried out inside an industrial oil and gas facility.
NOTE: Cognite Atlas AI is a low-code industrial AI agents workbench that powers agents with AI-ready industrial data to automate industrial workflows. It is a an industrial agent service that promises to revolutionise how industries approach AI-powered operations. The company says it has made it easy for anyone to build and deploy hundreds of reliable AI agents within minutes.
Unlocking the value chain
Looking to analyse how organisations can now work to unlock deeper value chains inside their business, CEO Rishi welcomed Kris Narayan from Nova Chemicals on stage. As the company’s data platform leader, Narayan explained how (however many use case datasets his team created) the resultant job would always lack the semantic context links (so that the data was really deeply linked to business operations, target and goals) that were needed for the organisation to really elevate and succeed. With Cognite’s enterprise knowledge graph, deployment meant that Nova would apply these data-driven advantages not just inside its plant, but also – wider – outward throughout its supply chain.
Analyst house Gartner has said that, “In the CIO race to deliver AI outcomes safely and at scale, the convergence of AI, data and digital transformation presents significant opportunities for Cognite, particularly in operational and enterprise-wide unstructured data applications.”
In an earlier blog post, Rishi has reminded us that Cognite’s expertise in operational data management and hybrid AI – combining generative AI with process and simulation models – gives the company a distinct edge. “Our ability to deliver scalable, repeatable solutions aligns us with the highest-value areas for AI innovation, ensuring we remain at the forefront of this transformative market. By focusing on our competitive edge and honing our core data modeling and AI capabilities, we will unlock new opportunities and deliver exceptional value as market dynamics evolve,” he said.
Birth of the industrial agent
“We have earned the right to operate across diverse verticals, partners and use cases. In 2025, we launched Cognite Embedded to empower OEMs and ISVs to accelerate product development by 2-3x and reduce total cost of ownership by up to 80%. System integrator partnerships with firms like Radix have generated millions in new pipeline opportunities. We reinforced our position as a cloud-agnostic leader in industrial data and AI solutions,” said Rishi.
“The power of Cognite Data Fusion and Atlas AI is redefining developer value creation. Our 100% API-driven software, coupled with a thriving online community for sharing best practices and code, puts innovation first. With reusable widgets and new hosting options for key partners, we’re enabling an ecosystem for rapid, scalable and secure application development,” said Chirayu Shah, CPO at Cognite.
Showcasing work carried out with Cognite in recent times and speaking at this year’s event was Paula Doyle, chief digital officer at AkerBP in line with Cognite’s exec VP Francois Laborie. Aker BP explores and produces oil and gas on the Norwegian continental shelf. Doyle herself has held a variety of roles within the oil and gas industry for companies such as ABB and Siemens in Norway and the Middle East. She has deep knowledge of the industrial software space and digitalisation processes in heavy-asset industries.
“We built a data foundation with Cognite at the heart of our operations,” said Doyle, speaking at the Impact 2025 in Houston. “Industrial data is multi-form, multifarious and massive (with legacy systems to also juggle), so we have worked alongside Cognite from its initial foundation as a business in 2016; our operational and engineering data is enriched with context and through Cognite Data Fusion before it is then fed into applications and agentic AI services. With its wider capabilities to also connect with data sources residing inside many different physical engineering components (everything from industrial pumps to compressors and more), Cognite enables us to operate at an outstanding 97% (or higher) efficiency rate.”
Aker BP’s “AI-first strategy” was developed to help lead the energy sector into a data-driven future by seamlessly integrating AI applications within industrial settings and business objectives to innovate and deliver results across the business. This includes automating complex processes and allowing engineers to easily access industrial expertise.
“Our collaboration with Cognite on Atlas AI represents a significant leap forward in our digital transformation journey,” said Paula Doyle, Chief Digital Officer, Aker BP. “This partnership underscores our commitment to a continued AI-first, data-driven approach with Cognite, embracing agentic AI to further drive efficiency, enhance decision-making and optimise our operations across the board.”
Cognite Atlas AI offers pre-built agents and a low-code workbench for building industrial AI agents, accelerating the journey towards autonomous operations. Integrated with the robust data foundation provided by Cognite Data Fusion, the collaboration supports Aker BP’s span of various strategic areas where AI is being leveraged to drive significant operational advancements and value.
During Cognite Impact 2025, the company announced a union with Databricks that focuses on “bidirectional zero-copy data sharing integration” between the two firms. The integration brings together Cognite Industrial AI and Data Platform, which incorporates Cognite Atlas AI and Cognite Data Fusion with the Databricks Data Intelligence Platform, including Databricks’ flagship AI product Agent Bricks.
What is bidirectional zero-copying?
A mouthful in its own right, bidirectional zero-copy data sharing can be explained as the practice and method of exchanging data between two applications, components, systems, processes or workflows without duplicating the information sets that reside in memory. It means that both parties can access (or modify) the shared data resource directly. All of which rather begs the question, what is industrial AI such hard work anyway?
Cognite points out that industrial AI solutions demand accurate real data with context, but this is tough given the widespread existence of complex, raw siloed industrial data. Cognite and Databricks promise to offer a new approach by using semantic models in the form of knowledge graphs with enriched context data to move beyond generic data models. This work aims to enable access to high-quality, trusted data required to build domain-specific AI agents and applications that reliably solve industrial challenges. It will (says Cognite) be able to eliminate expensive data duplication, storage and complex ETL pipelines.
Other news from Cognite at this event saw it announced that the Cognite Industrial AI and Data Platform, which includes Cognkite Atlas AI and Cognite Data Fusion, is now integrated with Nvidia accelerated computing and software. This focuses on real-time anomaly detection in time series data. This joint effort with Celanese, a global speciality materials and chemical company, aims to enhance the speed, scale and accuracy of identifying critical anomalies in vast streams of industrial time series data, enabling proactive intervention to help prevent operational disruptions.
In Cognite Atlas AI, a connection exists here to NV-Tesseract, a foundation model family developed by Nvidia for time-series analysis, purpose-built to rapidly process and extract insights from complex industrial data streams. The companny also released details of its partnership with Snowflake, where plans also provide a bidirectional, zero-copy data sharing integration between the Cognite Industrial AI and Data Platform and the Snowflake AI Data Cloud.
The integration will utilise a zero-copy data sharing integration to deliver a seamless, bidirectional flow of AI-ready data between the Cognite Industrial AI and Data Platform and the Snowflake AI Data Cloud. This enables Snowflake end-users across the enterprise to gain real-time access to unified, domain-specific industrial data needed to power AI solutions and agentic workflows, while simultaneously allowing insights derived by these users to continuously enrich the Cognite platform.
Has industrial AI come of age?
Technology vendors have been talking about the rise of so-called “industrial AI” for a couple of years now, but no formalised definition of what this would really mean in technical and operational terms has been laid down (spoiler alert: that story has to be told), so for now, we can take it from Cognite that this space is characterised by the fact that industrial operations generate an immense, continuous flow of sensor readings and equipment performance metrics.
These readings are often noisy, complex and lack context (essential background information and relationships), making it challenging to separate the predictable signal from random noise. Industrial AI could be the toolset and injection of galvanised grease that solves these riddles.
Secret CEO superpower?
Cognite CEO Girish Rishi is like a number of other Indian-born engineers who have progressed to impressive C-suite roles and have become fully absorbed into the Silicon Valley or wider North American technology glitterati. But, significantly, Rishi is unlike almost any other CEOs (and most C-suite technology execs) in that he actually got his hands dirty on the keyboard and performs the central product demos of the Cognite platform himself during his company’s keynote. If that’s his secret superpower ability that he pulls out at customer prospect sessions, then all power to him for being realistically capable.
… and (just in case you were wondering) what is the Cognite “moonshot” that CEO Rishi pledges to achieve? It’s a mission (and a promise) to deliver $100 billion of customer value by 2035. With some 1.5 billion customer “actions” being processed through Cognite per month, the source here is rich for sure.