Cognite conference rides crest of AI tsunami

Almost tripling last year’s attendance, Cognite’s Impact 2025 industrial and artificial intelligence (AI) conference sold out with 800 visitors during its Oct. 14-16 at the Houston Galleria. The three-day event featured multiple technical sessions, partnership announcements, customer case studies and exhibits highlighting how the company and its partners’ industrial AI expertise, along with its  Cognite Data Fusion (CDF), low-code Atlas AI and Canvas software, deliver benefits for users in multiple manufacturing sectors.

“We’re the only company founded on using large-language model (LLM) to drive value for you, but the world has changed a lot in the past year,” said Girish Rishi, CEO at Cognite, during his keynote opening Impact 2025. “Reindustrialization and reshoring that began at end of the COVID-19 pandemic are really gaining momentum, and now oil and gas are fashionable again. We’re also seeing for the first time a worldwide renaissance in industrial AI, which is also patriotic because it’s joined with national and cybersecurity. Cognite also serves that calling because users want data that secure in their companies and countries.”

Rishi reported that Cognite’s three primary goals are to unlock:

• Workforces with agentic AI and AI-ready, intelligent data in augmenting the industrial workforce;
• Value chains by extending access and value of AI and intelligent data from plant-floors to the enterprise levels and across supply chains; and
• Ecosystems with open platforms to encourage more AI-driven value across an expanded ecosystem of partners.

“For example, when operators learn they have excess oil dissipation in a compressor, they already rely on Cognite for graphs and contextualized data, and they have specifications and procedures for  equipment to also help diagnose causes. However, they can now use our Canvas software with our Atlas AI software agents to create cause maps,” explained Rishi. “This not only informs users about causes, but it also enables them to create notifications, establish preventive thresholds for better responses in the future, and even send workorders to SAP software.”

Users anticipate AI leaps

Kris Narayan, digital enablement and data platforms leader at Nova Chemicals in Calgary, Alberta, reports it previously drew data from its enterprise resource planning system (ERP) to create models, but lacked the semantics to process each data set into an insight-rich state until it adopted CDF software. This makes production information useful for more than 300 users, but it can also integrate supply chain and finance input from beyond the plant into knowledge and enterprise graphs. Nova implemented CDF at one plant, later expanded it to 10 assets, and now plans to roll it out at all of its facilities by the end of the year. 

“These maps can show processes and identify problem areas, but now we can ask Atlas AI to do root-causes analyses. This generates results right away, instead of the three months it used to take just to gather materials,” says Narayan. “The plant is where value is created, but now we can connect the dots to every related enterprise. Atlas AI agents can provide more details on batches and customers impacted, run executions, and develop remedies with suppliers.”

Karl Johnny Hersvik, CEO at Aker BP, added it’s relying on Cognite’s AI capabilities to help it liberate more data, make it more accessible, and operate three times more efficiently, on time and on budget, as well as discover more oil and gas resources. “Modern AI platforms must handle multiple data sources and types, and simulate thousands of productions runs, scenarios and outcomes before they occur, so they can guide decisions,” said Hersvik. “We launched our first generative AI implementation in 2024 to do document parsing, and this year we have AI agents helping with root-cause analyses. We want to out AI into the core of every problem we need to solve, and we’re expecting it to help increase our efficiency by 300%.”    

Value and partnerships

To advance these capabilities even further, Rishi added that Cognite also plans to deliver $100 billion of realized value to its customers by 2035. “This bold goal isn’t just an aspiration. It's a pledge of partnership,” he stated. “By unlocking the workforce, value chain, and the entire ecosystem through industrial AI, we’re committing to making that value a tangible reality for every customer.”

This goal will be aided by three partnerships launched at Impact 2025:

• Databricks will collaborate with Cognite on bidirectional, zero-copy data sharing between CDF, Atlas AI and Databricks’ Agent Bricks AI software. They’ll employ the open-ecosystem approach used by both companies to provide a unified, domain-specific, intelligent data foundation for industrial AI. They’ll use semantic models/knowledge graphs to provide enriched information/context about customers’ data from their production processes. This will allow them to move beyond generic data models, gain instant, real-time access to unified, domain-specific data for powering reliable AI and machine learning (ML) models, and enable reliable agentic AI.
• Nvidia accelerated computing and software, including its NV-Tesseract model family, will integrate with CDF and Atlas AI to provide data context, extend software agents into workflows, build AI models that understand manufacturing, and enable new levels of efficiency, intelligence and innovation for asset-heavy industries. They’ll initially work with chemical manufacturer Celanese on real-time anomaly detection in vast streams of time-series data (TSD), enabling proactive intervention to help prevent operational disruptions.
• Snowflake will also collaborate with Cognite to develop a bidirectional, zero-copy data sharing integration between CDF, Atlas AI and Snowflakes’s AI Data Cloud software to give end-users across enterprises real-time access to unified, domain-specific data needed to power AI solutions and reliable agentic AI workflows, while simultaneously allowing insights derived by those users to continuously enrich Cognite’s platform.