Google Gemini Enterprise Can Now Search BigQuery Data Without Copying It

Google Cloud has rolled out a major update for Gemini Enterprise that lets the AI model search a company's data directly in BigQuery, Cloud SQL and Spanner — without copying or moving that data into a separate AI store first. The zero-copy design uses the open-standard Model Context Protocol (MCP), and it means businesses can run powerful AI analytics on their private data while the data itself stays exactly where it already lives.
The key benefit is security. The AI only retrieves data based on the specific credentials of the employee making the request, so a company's existing access rules and hierarchies stay intact — a junior analyst cannot suddenly see files reserved for senior managers just because an AI is doing the searching. For firms in banking, healthcare and other regulated industries, that removes one of the biggest worries about adopting AI: the risk and cost of duplicating sensitive information into yet another system.
The move reflects a wider shift in enterprise AI. Rather than pulling company data into the AI's own warehouse, tools are increasingly being built to work inside the systems businesses already trust. For Google, it also strengthens Gemini Enterprise's pitch against rivals: companies get the analysis they want, with their data staying put, their compliance teams satisfied, and no extra copies floating around to secure.
Source: read more
Comments
Post a Comment