For a few month now, anyone with Data Studio access can ask questions of their data in plain language, which moves the decisive question from who has a licence to whether the data underneath can carry the answer. 

Conversational Analytics is the chat interface inside Data Studio that turns a typed question into a query and an answer. Since April 2026 it is available to all Data Studio users rather than only to Pro subscribers, and it no longer requires Gemini in Data Studio to be switched on; it has been generally available since July 2026.

In this form it talks to data agents that are built in BigQuery and published to Data Studio, so the people who will use it are usually not the people who configure it. It affects anyone in a marketing or analytics team who today waits for a report to answer a one-off question.

What's New

1. Open to all 

Access is no longer tied to a Data Studio Pro subscription, and the feature no longer requires Gemini in Data Studio to be enabled. That change is dated 16 April 2026 in the Data Studio release notes, with general availability on 30 July 2026. 

1. BigQuery agents 

The current experience connects to BigQuery data agents only. The older experience that also covered CSV files, Google Sheets and Looker sources is being phased out, and existing conversations and agents cannot be migrated across. 

1. Built upstream 

Agents are not created in Data Studio. They are built in BigQuery, given system instructions, table and column descriptions and example queries, then published to Data Studio. Access follows BigQuery permissions rather than report sharing.

1. Advanced stays paid 

The Code Interpreter, which handles cohort and retention analysis, outlier detection, growth rates and time series work in Python, still requires a Data Studio Pro subscription with Gemini in Data Studio enabled.

What you can do with it 

• Answer routine questions about campaign, channel and on-site performance without commissioning a new report for each one. 

• One governed definition: a single published agent means the logic behind an answer is maintained in one place, rather than being re-derived in every dashboard. 

• Review how an answer was produced. Since February 2026 the chat shows its reasoning step by step, so the generated query can be checked rather than taken on trust. 

• Keep cost predictable. Query spend can be capped per request, and agent usage, latency and token consumption can be monitored centrally. 

Current limitations 

Data sources: The open experience reaches BigQuery data agents only. Data held in Sheets, CSV files or Looker Explores is not covered, and legacy conversations cannot be carried over. 

Where agents are built: Agents are created and maintained in BigQuery, not in Data Studio, so setup depends on warehouse access and on written context that someone has to keep current. 

Advanced analysis: Forecasting, cohort analysis and similar work require the Code Interpreter on a Pro subscription. Looker queries there are capped at 5,000 rows and map visualisations are not supported. 

On the roadmap: Google continues to ship on the platform side: agent monitoring and HIPAA compliance arrived in September 2026 and BigQuery Graph support reached general availability on 30 September 2026. A deeper multi-step reasoning mode is in limited preview with no published release date yet.

What we recommend 

Use it for ad hoc questions: It is a good fit today where one well modelled BigQuery dataset answers recurring one-off questions, and where someone owns the agent context. Start with a single scoped agent rather than opening it across the warehouse. 

Keep reporting as it is: Your governed dashboards and reporting layer stay unchanged. Conversational Analytics answers questions against one agent; it does not replace defined KPIs, scheduled reporting or a modelled data layer, and Google notes its output can be plausible and still wrong.