Artificial intelligence (AI) has moved past the initial stages of the gold rush. In a scramble to avoid getting lost in the noise surrounding the generative AI hyperbole, IT suppliers hurried to extol the virtues of their data-enabled products. Now, following that first stage of the gold rush, technology companies are focused squarely on stage two – becoming the trusted platform for AI.
Into this competitive marketplace comes data specialist Snowflake, a company with a long history of producing enterprise-ready IT platforms for blue-chip enterprises. At its Snowflake Summit 2026 event in San Francisco, the firm’s senior executives outlined how the firm is pivoting strategically to meet customers’ demands in the age of agentic AI.
“Last year we could not have imagined some of the things that we are doing today,” said Snowflake CEO Sridhar Ramaswamy, outlining the pace of change at the beginning of the conference’s keynote presentation. “AI isn’t a promise anymore; it’s creating real outcomes, real tangible opportunities, and our goal this week is to make AI real for your business.”
How is Snowflake planning to pivot to help its customers meet their objectives as the pace of transformation quickens? The answer, said Ramaswamy, is that the company wants to create what it calls an agentic control plane for AI. This orchestration layer will connect the AI models and services that customers use to exploit data in their Snowflake platforms.
“You can spin up agents across your business, but without a control plane, these agents are effective only in isolation and not really aware of each other,” said Ramaswamy in the keynote. “You need a way to coordinate across your context models and applications, so that decisions and actions happen seamlessly across your business.”
On the surface, Snowflake’s pivot to become a platform for AI services is similar to shifts by other technology providers. Computer Weekly heard in Dublin earlier this year how Workday aims to provide an agentic AI platform that disrupts traditional enterprise resource planning services. In New York City, meanwhile, we heard that Freshworks wants to create an agile AI-enabled platform that connects assets and incidents. What’s clear from these transformations is that technology companies must find a way to establish a presence in the fast-evolving world of enterprise IT. The first keyword for these transitions is “platform”. While Workday is focused on employee experiences and Freshworks is centred on customer support, Snowflake aims to become the trusted platform for the agentic enterprise, where customers securely exploit structured and unstructured data.
The second keyword is “open”. In an age when no one can be certain which models will dominate today, never mind tomorrow, providers are eager to give their customers a choice. Snowflake executives explained that its agentic control plane will use application programming interfaces and Anthropic’s Model Context Protocol to ensure flexibility, whether interacting with other enterprise applications, such as Salesforce and Workday, or the frontier AI models.
“The notion that you need to put all your data in one platform and decide everything with that one vendor is a tale of the past,” said Snowflake’s executive vice-president of product, Christian Kleinerman, at the event. “What we’re offering is choice. If you already have data outside, you can use Snowflake to make reads or writes to that data.”
Pivoting strategically
So, is Snowflake’s fresh focus on openness a significant shift in approach? When Computer Weekly suggested to Snowflake’s co-founder, Benoît Dageville, in a one-to-one chat that some people might have seen the firm’s previous approach as more closed, he was eager to suggest that interoperability has always been a core element of Snowflake’s strategy.
“We created the Marketplace so that you can interact with data,” he said, referring to the firm’s long-established hub that allows customers to buy third-party datasets, services and applications. “A lot of things have been said about Snowflake being closed. It’s not that we were closed; we invented this technology, the data platform. And because we invented it, there was no one to interact with.”
Having built and honed its data platform, the company’s aim in the agentic AI era, said Dageville, is to allow CIOs and other business leaders to run models and services on its data foundations. At a time when many providers fear being disintermediated, Snowflake wants to ensure it remains at the core of whatever long-term shape enterprise technology assumes.
“If you want to create the control plane for AI, you need it to interact with any aspect of your enterprise,” he says. “Whatever you use in your business, you should be able to connect to agents, so interoperability is a critical success factor. Choice is a very important aspect. Snowflake gives you the choice, because we can run every model.”
Interestingly, Snowflake’s open approach also includes some in-house innovation. In addition to allowing customers to hook other models and software into its data platform, the technology company continues to develop agents as part of its agentic control plane.
At the San Francisco event, Snowflake announced new features and a rebrand of Snowflake Intelligence, its personal agent for knowledge workers (now known as CoWork), and Snowflake Cortex Code, its coding agent for developers and data engineers (now known as Snowflake CoCo). Dageville said this broad approach, with in-house-developed agents and external models, helps Snowflake to support choice.
“We are living in the age of the agentic enterprise,” he said. “Your platform will communicate with a set of AI agents, whether these agents run inside Snowflake or outside with MCP. The control plane is critical because of the governance and the data that it can leverage.”
To this end, Snowflake announced additional features at the event, including Datastream, a fully managed streaming service to ingest real-time data into the Snowflake platform, and Horizon Context, a governed semantic foundation that ensures every person, tool and AI agent operates from the same source of information. During a media panel, Ramaswamy put the firm’s agentic developments and feature updates into context.
“Our thesis is that the data platform as a source of trusted, secure, governed data will continue to be important, and in fact, what you can do with a data platform like Snowflake is going to amplify, get larger, get easier, and that’s a theme of the summit,” he said. “It’s the culmination of many quarters of work the Snowflake teams have put in.
“One set of features is around core data platform changes – new features that are going to make our platform even more powerful,” said Ramaswamy. “But a lot of the summit’s theme is about how you use agentic AI to get things done faster with Snowflake via coding agents like Coco, and how you can begin to think about work very differently with a tool like CoWork.”
Pushing boundaries
Perhaps the most intriguing element of Snowflake’s AI vision is the firm’s commitment to developing its own agents. Yes, the company is opening its platform to frontier models from innovators such as Anthropic and OpenAI, but the firm’s in-house agentic developments suggest it believes it can provide something, if not better, then certainly different, to the models being created by the LLM specialists.
Responding to a question from Computer Weekly during the media panel, Ramaswamy said Snowflake must continue to innovate in all areas of AI. He suggested the development work the company is undertaking with its agents now will allow it to push beyond new boundaries in the future.
“Anyone who is sitting out how users are going to interact with software is putting themselves in great strategic peril, and there is also a world of learning that comes from creating these products, both internally, but definitely with your customers,” he said.
That engaged approach was confirmed by Kleinerman, who suggested the firm’s agentic approach is a strategic pivot. “Snowflake used to be primarily a back-end specialist,” he said. “Our work on agents is an example of how AI is changing things in a way that allows us to rethink that approach.”
So far, the market likes the direction of the firm’s pivot. While some IT companies have struggled to post impressive figures in challenging market conditions, Snowflake’s latest quarterly results, released just before the event, show total revenue reached $1.39bn, representing a 33% year-over-year increase. The impact of the strategic pivot the company highlighted in San Francisco will be seen in the next set of results. However, Computer Weekly also spent time with customers at the event to elucidate their opinions. The general conclusion from the digital leaders at these businesses was that they like what they’re seeing.
Agentic services
Stephen Wood, chief operating officer at Rathbones Asset Management, said he’s interested in the opportunity to exploit Snowflake’s agentic services. He said services like CoCo and CoWork show how his firm could bolster its resource-constrained engineering resources with specialist technological capability.
“If there’s a future scenario where you’ve got an agentic service and a model that sits on top of the cloud platform that you’ve already invested in, and you can leverage that technology, bring it into your business, and train it on the 20% that is relevant to you, that’s incredibly powerful for many, many businesses,” he said.
“That’s because the time to market is really quick. This speed means you can suddenly access technology that you ordinarily wouldn’t have. And that potential means the capability closes the gap between the bigger players and us. We need to be able to continue to leverage this type of technology quickly, and I think service models that sit natively in the infrastructure that you already have are the way to do it.”
That sentiment resonated with Joe Zhang, chief technology officer at specialist healthcare partnership OneLondon, which connects health and care information across the capital to improve patient treatments and plan future medical services. Zhang said the potential strengths of Snowflake’s technology are clear, but must be considered in a wider context.
“Technology, and what we can do with data, and also the culture around data, has moved so much over the past five years,” he said. “While much of this technology is new, we remain at risk in healthcare of being outpaced by everyone else. We always have been, and the pace of change itself is increasing now.
“The agentic part has extraordinary potential. There are a lot of workflow inefficiencies in the NHS that could be mitigated by the technology. But there are also higher risks than in many other industries, so it’s about treading carefully and balancing risk versus potential gains.”
Leave a Reply