Snowflake CoCo lands in Microsoft Excel and Google Sheets in public preview

Snowflake's move addresses a long-standing gap: spreadsheets are where many business analysts spend their time, but connecting them to governed cloud data has usually meant manual CSV exports or asking IT to set up ODBC or third-party connectors. Those workarounds create disconnected copies and leave administrators struggling to keep versions straight. By placing CoCo directly in Excel and Sheets, Snowflake aims to let users work with live data without leaving the spreadsheet.

The company says it tested CoCo for Excel against an AI model harness for Excel using the Snowflake MCP across 25 spreadsheet tasks with Opus 5.5, Sonnet 5.5 and Haiku 5.5. The add-ins had similar accuracy, but CoCo for Excel showed better latency and used fewer output tokens. Snowflake also reports near-constant performance as data size scales, which it attributes to streaming query results directly to the sheet instead of through the model's context window. These internal results have not been independently verified.

Customers cited in the announcement describe easier handling of complex tables. Matt Wyman, Vice President of Information Technology at The Chefs' Warehouse, said CoCo understands table relationships and is intuitive enough for analysts to use without training. Kate Savina, Head of Data Technology at GTCR LLC, said CoCo starts with the firm's data directly, with no exporting or switching tools. Snowflake says both add-ins are in public preview now.