Google's Gemini can now answer questions from your company data

Google has added a way for employees to ask Gemini a question in everyday words and get an answer from live company data. The feature is called Data Cloud connectors in Gemini for business, and it is now in preview. It aims at a familiar problem: data questions pile up faster than analysts can answer them.

What changed

Until now, a manager who wanted a number often had to file a request and wait. With the connectors, that person types a question into Gemini. Google's example is, "How many sales did we have this month for all cold weather products?" Gemini finds the right data, writes the database query and runs it. Users do not need to know which table or project holds the data.

Results can then move into everyday tools. Google says they can become a Google Doc, an email draft in Gmail or a deck in Google Slides.

How it works

Gemini works with Knowledge Catalog, a Google tool that keeps track of what data a company has and what it means. It turns everyday terms into exact data labels. In Google's example, "cold weather products" becomes categories such as sweaters, coats and jackets. The flow has four steps:

• You ask a question in plain language.
• Gemini searches Knowledge Catalog for likely data.
• Business terms are matched to the exact data structures.
• Gemini writes the query and runs it under your login.

Who can see what

Queries run under the identity of the person asking. Existing row-level and column-level permissions still apply, so people only see the rows and columns they already have rights to. Google says an employee without access to executive salary data or protected healthcare records cannot get it by asking Gemini. Gemini also only discovers catalog information that the person is allowed to see.

Where the data comes from

The connectors cover BigQuery, Spanner, AlloyDB, Cloud SQL and Looker. All are in preview. The BigQuery connector can also reach open file formats such as Apache Iceberg, Delta Lake and Parquet. It can reach datasets in AWS or Azure as well.

Google also offers Conversational Analytics agents. These are narrower assistants that a company's data team builds and certifies for a specific use, aimed at governed, repeatable reporting. Gemini can point users to published agents, so they can choose between open-ended questions and a scoped agent.

Setup and open questions

Administrators turn the connectors on in the Gemini admin console. They set up default read-only connections to BigQuery, operational databases or Looker, and Knowledge Catalog is enabled automatically. Users then switch the connector on in Gemini.

The announcement does not give pricing or a date for wider release. It also offers no accuracy figures, which matters because earlier AI data tools often struggled with complex data.