Snowflake outlines what businesses need to become agentic enterprises

At its Snowflake World Tour London event, Snowflake leaders described the foundations they say businesses need to embed AI agents in core operations. The company’s view centers on combining AI models, business data and software in what it calls an “agentic control plane.”

James Hall, Snowflake’s Country Manager for UK&I, said there is no enterprise AI strategy without a data strategy. He called for a data foundation that is trusted, governed, secure and accessible, and said businesses had moved from discussing AI’s potential to putting it to work in the past 12 months.

Christian Kleinerman, Snowflake’s EVP of Product Management, described an agentic enterprise as an organization where AI agents work within core processes, coordinate with people, follow direction and guardrails, and automate tasks. He said the aim is to help workers become more productive and improve business outcomes.

Kleinerman identified three elements: an AI model, data context, and software and apps. He said Snowflake is committed to offering customers a choice of models, since it is unclear which models will lead even in the next few months. He also argued that useful AI depends on understanding and governing business data, while software and apps connect AI to enterprise infrastructure.

The article quotes Kleinerman calling the control plane “the glue” that enables models to interoperate with that infrastructure. Hall said moving quickly with data and AI should not mean losing control. He described the agentic enterprise as already happening in production.