AirQuery is a fully managed cloud platform for conversational AI analytics. Connect it to Databricks in minutes, with nothing to install. Your experts build Thinkboxes on Databricks tables, you pick the model, and business users ask questions in plain language. Every answer is reproducible, cited, and governed by the security rules in your AirQuery data model. Your existing Unity Catalog row filters and column masks are supported too.
EMEA operating margin fell 3.4 pts to 18.1%. Most of the drop came from higher freight costs in two distribution centers and a price-mix shift toward lower-margin SKUs.
Sign up for AirQuery and connect Databricks from the AirQuery console. Connect with a Databricks service principal or OAuth. AirQuery uses a read-only connection you control, reads table metadata, and sends queries to your Databricks SQL warehouse, so your tables are never copied into AirQuery.
Your subject matter experts pick the catalogs, schemas, and tables that matter. Then they define relationships, approved metrics, business rules, and security rules for who can see which rows and columns. Thinkbox assembles these into a living model the agent reasons over.
Users ask questions in the AirQuery web workspace, Slack, or Microsoft Teams. Each answer shows its sources, SQL, and reasoning steps. Pin answers, set alerts, or embed them in your own apps through the API or MCP.
Databricks gives you one platform for data and AI. AirQuery adds a conversational AI layer on top, delivered as a cloud service, so business users get answers instead of tickets. Thinkbox combines model intelligence with the knowledge of your own experts. The result is analytics your CFO can trust and your analysts don't have to rebuild.
AirQuery's cloud reasoning layer sends SQL to your Databricks SQL warehouse. The heavy lifting happens in Databricks, with no extracts, no copies, and no second warehouse to maintain.
Use AirQuery's managed models, or connect models you already run in Databricks Model Serving. Change models without rebuilding your Thinkboxes.
Row-level and column-level security rules live in the Thinkbox data model, next to your metrics. Define them once and they apply to every question, in every channel. Already use Unity Catalog? AirQuery works with your existing policies as well.
AirQuery is a SaaS platform built for finance, operations, and regulated teams, where a wrong number or a leaked row is not an option.
Security rules are part of the AirQuery data model. Map users and groups from your SSO to the rows they can see and the columns they can access, and mask sensitive fields such as salaries or personal data.
AirQuery supports Unity Catalog row filters and column masks. Each query runs with the asking user's context, so your existing policies filter rows and mask columns inside the database, on top of the rules in the AirQuery data model. The connection itself uses a read-only account you control.
AirQuery sends SQL to Databricks and receives only the result rows needed for each answer. It never replicates your tables, and your data is never used to train models.
Every answer records the user, question, model, Thinkbox version, security rules applied, generated SQL, and source tables. Re-run it and you'll get the same answer.
The same metrics and security rules apply whether a question comes from the web workspace, Slack, Teams, the API, or MCP. There are no per-dashboard permissions to maintain.
SOC 2 Type II, SSO, role-based access to Thinkboxes, encrypted connections to Databricks, and encryption in transit and at rest.
Most Databricks projects end up with clean data and a long backlog of report requests. AirQuery opens that data to the people who need it, with the same guardrails your data team built.
No servers to deploy and no upgrades to manage. Connect Databricks, build a Thinkbox, and business users can start asking questions the same day.
Deterministic reasoning over approved metrics removes hallucinated numbers and conflicting definitions.
Security rules live in one place, the AirQuery data model, and are applied to every answer automatically. No shadow copies of your data.
Turn your Databricks investment into everyday decisions across finance, sales, supply chain, and HR.
"What's driving the gap between forecast and actuals this quarter?"
gl_entries × forecasts × deals · main.finance
"Which suppliers caused the most late deliveries in the last 90 days?"
purchase_orders × receipts × suppliers · main.scm
"Which enterprise accounts are at risk of not renewing next quarter?"
subscriptions × usage × support_tickets · main.revops
"How has attrition changed by department since the reorg?"
workers × assignments × terminations · main.hr (masked PII)
Connect Databricks to AirQuery and see governed answers on your own data. Start free with a 14-day enterprise trial, no credit card required.