AI for Database: Natural-Language Database Analysis

AI for Database lets a team ask questions about a connected database or spreadsheet in plain language and inspect the resulting SQL, charts and dashboards. The platform also describes scheduled workflows and alerts. Answering a question and executing an action are separate capabilities with different permission needs.

Visit AI for Database

Sources checked October 27, 2026 · Public-source review; no hands-on product test

Ask, inspect and turn an answer into a report

The official product page describes natural-language questions against connected data, SQL inspection, charts and dashboards. It also lists alerts and actions through workflows. A useful first question has an answer you can independently reproduce from the same data and date range.

Start with sample data or a suitably restricted connection. Inspect joins, filters, time zones and aggregation before sharing a result. A fluent explanation does not establish that the query used the intended business definition.

Per-user subscriptions fund individual credits

The pricing page shows Pro at $10 per user/month and Team at $30 per user/month, billed monthly. Each payment becomes the same amount of individual usage credit; credits do not expire. AI usage pauses when the balance runs out, and top-ups are available.

Annual billing is offered separately. Workflow creation and database connection allowances do not mean unlimited AI compute. A free starting experience or sample dataset should not be described as a permanently free managed-AI balance.

Queries and actions need different safeguards

Use read-only access for an initial analysis task. Before enabling a schedule, webhook or connected action, review what it can read or change and who can approve it. A natural-language setup does not remove the underlying database and account permissions.

This can suit a team repeatedly asking operational questions. A conventional SQL editor may be a better fit when an analyst already has a precise query and needs full manual control.

Before connecting production data

Read the official documentation, verify a known answer with sample data, and inspect the usage charge. We have not connected a database, run a query or tested an automated action. Security, deployment and enterprise commitments should be confirmed in the relevant contract.