Google Cloud launches agentic data tools for AI workflows

Google Cloud is expanding its Agentic Data Cloud with a set of tools intended to bridge the gap between AI models and enterprise data systems. The updates focus on enabling natural language interactions, automating data workflows, and improving governance across operational and analytical platforms. These changes target developers, data scientists, and business analysts who rely on real-time data for decision-making.

What was announced

The latest additions include conversational analytics tools integrated into BigQuery, AlloyDB, Spanner, and Cloud SQL. These tools allow users to query databases using natural language, eliminating the need for manual SQL writing. For example, BigQuery Studio now embeds an AI reasoning engine that grounds responses in business context, while Lakehouse support enables cross-cloud queries without data movement. Looker’s embedded conversational analytics, now generally available, lets organizations embed AI agents into custom applications via low-code iframes.

Google also introduced several specialized data agents. The Data Engineering Agent, now generally available, automates pipeline creation and maintenance by converting natural language requirements into optimized SQL or Python code. The Data Science Agent, currently in preview, assists with feature selection, notebook code generation, and documentation. Database-focused agents, such as the Database Observability Agent and Database Onboarding Agent, monitor performance and recommend deployment configurations, respectively. These agents are designed to reduce manual intervention in data management tasks.

For developers, Google released tools like the Data Agent Kit, which provides standardized skills for agentic development within IDEs and CLIs. Managed MCP Servers for databases and Looker simplify secure connections between AI models and enterprise data, while QueryData converts natural language into database queries with near-perfect accuracy. The Universal Commerce Protocol (UCP) Analytics integration streams real-time commerce events into BigQuery for observability in agentic workflows.

Why it matters for enterprises

The updates address two persistent challenges in AI adoption: the lack of enterprise context in generic AI platforms and the security risks of ungoverned data access. By embedding AI reasoning directly into databases and analytics tools, Google aims to improve accuracy and reduce hallucinations. The agents also enforce granular access controls, ensuring compliance with enterprise governance policies. For example, the Deep Research Agent synthesizes intelligence from internal documents and public sources while respecting user permissions.

Business users benefit from simplified access to data insights. Conversational Analytics in Gemini Enterprise allows non-technical teams to interact with agents built in BigQuery or Looker without accessing technical consoles. The Looker Dashboard Agent provides AI-generated summaries of dashboards, making data more accessible to decision-makers. These features could reduce the dependency on data teams for routine queries and reporting.

What to watch

Most of the new tools are in preview, with general availability limited to a few components like the Data Engineering Agent and Managed MCP Servers. Enterprises will need to evaluate the accuracy and reliability of these agents in production environments, particularly for high-stakes use cases. The cross-cloud capabilities of Lakehouse conversational analytics could also face scrutiny over performance and data residency concerns. As Google expands these tools, competition with other hyperscalers offering similar agentic data solutions will likely intensify.