Snowflake has announced the renaming of two of its AI offerings and unveiled a series of new capabilities spanning AI, interoperability, security and data streaming ahead of Snowflake Summit 2026. “The future of enterprise AI will be defined by how well organizations connect intelligence, trusted data, and action across the business,” said Vijayant Rai, Managing Director- India, Snowflake. “At Snowflake Summit 26, we’re introducing innovations that help power the agentic enterprise, giving organizations a trusted foundation and control plane to build AI faster, operationalize it securely at scale, and enable teams and AI agents to work together from a shared business context wherever data lives.”
Addressing media during a virtual briefing, Christian Kleinerman, Executive Vice President of Product at Snowflake, said the company is rebranding Snowflake Intelligence as Snowflake CoWork and Snowflake Cortex Code as Snowflake CoCo. “Before going into our vision and announcements, I want to lead with an update that we will be sharing with our customers. CoWork is the new name for what we have been calling Snowflake Intelligence, which is the personal agent where we help organisations work smarter. It is targeted at non-technical workers and specifically knowledge workers,” Kleinerman said.
“The other change that we’re introducing is the renaming of Snowflake Cortex Code to Snowflake CoCo. This is the coding agent that helps individuals build faster and is targeted for a builder audience and a technical audience,” he added. Explaining the rationale behind the change, Kleinerman said customer adoption influenced the decision.
“When we introduced Snowflake Cortex Code, we affectionately called it CoCo. The excitement that we have seen from our customers about CoCo is extremely high. We realised that in CoCo, we have a very strong brand and decided to stop calling it Cortex Code and just adopt CoCo because our customers and our users have largely spoken.” He said the renaming aligns the products more closely while keeping their purpose unchanged. “To make it more in line, we decided to rename Snowflake Intelligence as CoWork. I mentioned this upfront because the rest of this presentation uses the names CoCo and CoWork.”
Snowflake Bets on the Agentic Enterprise
Kleinerman described the company’s broader vision around what it calls the “agentic enterprise”. “Maybe the most important message for all of you today and for our customers is how we think about the agent enterprise. “We like to think of the agent enterprise as an organisation where every single person is more productive, more efficient and more leveraged because they can leverage AI. Snowflake is the platform that can help any organisation transform how each individual does work and is able to leverage AI” he said.
He emphasised that enterprise AI must go beyond standalone assistants: “A very important distinction is enterprise context, with access to systems, access to data, but with the peace of mind of governance, compliance and security.” Highlighting Snowflake’s approach, Kleinerman said the company sees enterprise data as the most valuable and stable asset in an AI-driven world. “AI models keep changing and capabilities keep advancing. One model is better today, another one is better six months from now and another may be better a year from now. But the data is constant for customers. Our offer to customers is that we help AI models connect to enterprise data and enterprise context with security in mind,” stated Kleinerman.
He added that Snowflake’s architecture is designed to support connectivity across systems: “We want to make sure there is connectivity and interoperability to other business systems and other data that may not necessarily be managed by Snowflake. All of this is coordinated by what we call a control plane, which is either CoCo or CoWork.”
Fastest Product Adoption in Snowflake’s History
Snowflake said CoWork and CoCo are witnessing some of the strongest adoption rates in the company’s history. CoWork, formerly Snowflake Intelligence, has experienced what Snowflake describes as “the fastest ramp in product adoption in company history.” The growth trajectory has been steep. In Q3 FY26, approximately 1,200 customers were using Snowflake Intelligence. By Q4 FY26, adoption had climbed to more than 2,500 accounts, nearly doubling quarter-on-quarter. In Q1 FY27, customer adoption more than doubled again.
“The momentum on both CoWork and CoCo is super strong. The number of customers that we have is quite strong – 13,900 customers and growing, including 810 Global 2000 customers. Nearly everyone, 13,600 customers, is using our AI capabilities on a weekly basis,” he observed while also adding that Snowflake currently serves 13,912 customers globally. CoCo has witnessed similarly strong growth since launching in November 2025. Adoption grew from 4,400 customer accounts in Q4 FY26 to more than 7,100 accounts in Q1 FY27. Snowflake says that adoption of CoWork and CoCo is contributing to increased platform consumption and played a role in the company raising its FY27 growth guidance from 27 percent to 31 percent year-on-year.
Customers Report Measurable Business Outcomes
Snowflake shared several examples of how customers are using CoWork to improve productivity, accelerate decision-making and automate business processes. Some of the most notable examples are as follows:
- TS Imagine has built an AI agent that performs work equivalent to 8.5 full-time employees, helping users automate customer case resolution, analyse data and make faster trading and risk management decisions.
- Toyota Motor Europe reduced AI agent deployment timelines from months to weeks, while Samsung used CoWork during the Galaxy S26 launch to gain real-time visibility into customer response across regions and channels, allowing teams to adjust marketing spend, inventory and pricing while the launch was underway.
- Synopsys has deployed more than 20 AI agents across finance, legal, revenue operations, product management and IT teams. United Rentals has enabled more than 5,000 frontline employees across its branch network to query company data using natural language and receive immediate answers.
- WHOOP reported that access to enterprise data, which previously required specialist analysts and manual requests, is now available to hundreds of employees in real time.
Snowflake also disclosed internal productivity gains from using CoWork. The company said creating job descriptions, which previously took between 60 and 120 minutes, can now be completed in five to 15 minutes. Employee survey comments that once required weeks to analyse can now be processed within minutes.
Within Snowflake’s service delivery organisation, customer projects are being completed up to five times faster. Response accuracy has improved by more than 25 percent, implementation cycles have been compressed from days to hours, project margins have increased by 40 to 50 percent and customers are going live more than 40 percent faster. The company’s sales organisation estimates that AI agent capabilities could recoup productivity equivalent to approximately 90 full-time engineers annually.
CoCo Delivers Developer Productivity Gains
Snowflake also highlighted significant customer benefits from CoCo, its AI-powered coding assistant:
- Consulting firm evolv reported saving more than 500 hours of work in the first 20 days of deployment, generating an estimated US$100,000 in value. “Twenty days, 21,000 operations, or 600 hours of work delivered. That is sixteen work weeks compressed into less than a month,” the company reported.
- Infinite Lambda used CoCo to build a complete Customer 360 application in just five hours, bringing together customer data, churn insights, recommended actions and dashboards into a single experience.
- Fanatics reported reducing development timelines from weeks to hours, while Thomson Reuters said teams are delivering insights in days instead of weeks while modernising legacy systems and managing AI pipelines across more than 37,500 governed tables and 350 data sources.
Snowflake’s own engineering teams have also reported measurable gains from CoCo. The company said case resolution times have improved by 25 percent, engineer throughput has increased by 25 percent, complex issue resolution times have fallen by nearly 30 percent and engineering effort per support ticket has declined by around 40 percent. Snowflake further reported that developer productivity has doubled, measured through pull requests and code output per engineer, while more than 100 workflows across finance, HR, marketing and sales have been automated within weeks. According to Snowflake, CoCo outperforms leading coding assistants when operating inside the Snowflake environment and delivers up to 3.5 times better performance than alternative approaches using external integrations.
As Enterprises Struggle with Data Silos, How Is Snowflake Turning Interoperability into a Practical Reality?
Interoperability emerged as a major theme during the briefing, with Snowflake announcing support for additional Apache Iceberg capabilities, integration with Apache Iceberg REST APIs and the ability for Snowflake to provide storage for Apache Iceberg tables. Responding to a question from Tech Achieve Media on interoperability, Kleinerman said Snowflake is taking a practical rather than theoretical approach: “There are many companies that talk about interoperability, but they don’t really mean it. We are not only adopting Apache Iceberg and Apache Polaris as core standards. We have a number of members on the steering committee guiding the effort of Iceberg so that it is truly interoperable.”
He said the company has been working with multiple ecosystem partners to validate interoperability in real-world environments: “We are working with a number of industry players to validate that our technologies work well together. For example, with Microsoft Fabric, we have leveraged Iceberg to show that Fabric can read into Snowflake and Snowflake can read into Fabric and vice versa. So it’s real testing.”
Kleinerman also highlighted the role of CoCo in simplifying interoperability tasks: “The most interesting thing that we’ve done is we’ve added skills to CoCo to make sure that creating an Iceberg table and interoperating between catalogues is a very simple operation. We are not only making it possible, but we’re also making it easy.”
New Data Streaming Platform
Snowflake also announced Snowflake Data Stream, a native streaming solution fully integrated into the platform. “This is a native streaming solution fully integrated into Snowflake,” Kleinerman said. “It is compatible with Kafka, enabling customers to leverage existing applications and existing clients while streaming data into Snowflake-managed infrastructure.” He added that the platform allows streaming and analytics to coexist. “We enable streaming and analytics to live side by side by allowing topics to be materialised into tables in Snowflake. From the few customers that we’ve validated this idea with, we found extremely strong interest.”
Bringing More Context to Enterprise Data
Another major announcement was Horizon Context, an extension of Snowflake’s Horizon Catalog: “Horizon Context represents a set of capabilities that will help organisations better manage semantics and metadata about their data,” Kleinerman said. The capability incorporates technology from Select Star, a company acquired by Snowflake last year. “It will let us extract semantic information from business intelligence tools, relational databases and data transformation systems. The continued innovation around semantic views will help capture richer relationships and richer information that models the business for the data,” he added. According to Kleinerman, this additional context is critical for AI: “It is what enables AI, and in particular CoCo and CoWork, to make sense out of that data.”
Security for the AI Era
Snowflake also announced a range of AI-focused security enhancements. “The first announcement is a new AI security package that will scan for common risks and potential exposures, especially in the use of AI or agents, and flag those vulnerabilities to customers,” Kleinerman said. The company is also introducing new data exfiltration policies. “Customers will be able to say that if an agent is running an operation, certain data movements can be blocked,” he commented.
Snowflake is additionally rolling out multi-party approvals for sensitive actions. “For certain sensitive operations, it can be configured that two administrators or two users need to take action. Overall, we’re focusing on security for AI,” he said.
New Features for CoWork and CoCo
Snowflake announced several enhancements for CoWork and CoCo. “We’ll be introducing a personal work agent where each individual in a company will have their own set of skills, their own set of MCP connectors and, most importantly, their own memory and state,” Kleinerman said. The company is also introducing automation, scheduled operations and next-generation dashboards. “It’s not just about creating dashboards. It will enable collaboration, commenting, sharing and saving,” reiterated Kleinerman.
Snowflake also unveiled Cortex Sense, a capability designed to generate additional business context automatically. For CoCo, Snowflake announced broader availability through new form factors. “We are introducing an Excel plug-in, a Cloud Code marketplace plug-in, a Slack integration and, of course, our desktop application,” Kleinerman said.
