Skip to main content
Use Case

Ask your data in plain English

Business teams shouldn't need a SQL ticket to get an answer. Golden Analytics translates natural language questions into warehouse queries and returns attribution-aware results, no analyst required.

Natural language query
"Why did MRR drop last week in the East region?"
MRR fell 6.1% in the East region last week. The primary driver was a 19% increase in churn among mid-market accounts, concentrated in the SaaS vertical. New business partially offset this at +2.3%.
Mid-market churn
-8.4%
New business
+2.3%
The problem

Business questions don't fit SQL ticket queues

Data requests travel through a translation layer: stakeholder question to analyst ticket to SQL query to dashboard filter. Every step adds latency and loses context.

Days of latency per question

A business stakeholder with a metric question submits a ticket and waits. By the time the answer comes back, the decision window has passed and the context has shifted.

Translation errors accumulate

What the stakeholder asked and what the analyst coded are often subtly different. Metric definitions, filter logic, and date ranges get interpreted rather than specified.

Analyst time on low-value queries

Repeat questions consume senior analyst time on work that has no strategic value. The same segment filter applied to the same metric shouldn't require a human every time.

How it works

From question to warehouse query in one step

01

Type your question

Any team member types a metric question in plain English. The system understands business terminology mapped to your specific metric definitions and schema.

02

Query generates and executes

Golden Analytics translates the question into SQL, validates it against your semantic layer, executes it against your warehouse, and applies the attribution model to the result set.

03

Plain-language answer with drivers

The result comes back as a written explanation with a segment-level breakdown, not a raw table. Follow-up questions refine the analysis in the same conversational thread.

Capabilities

What NL Querying gives your organization

Semantic layer integration

NL queries map to your dbt metrics or custom semantic definitions. The system uses your canonical metric logic, not a guess at your schema.

Conversational follow-ups

Ask "what about enterprise accounts only?" and the system carries context from the previous answer. Each follow-up refines the same attribution analysis.

Row-level security respected

Queries execute under the same service account permissions as your other warehouse reads. Users can only see the data their warehouse policies permit.

Generated SQL visible

Analysts can inspect the exact SQL generated for every question. No black-box answers. Verify the translation against your own expectations and flag discrepancies.

Shareable query links

Share a permalink to any question-and-answer pair with your team. The link re-runs live against current data, not a cached snapshot.

Slack and API delivery

Receive answers in Slack channels or pull them programmatically via API. Embed NL query results in the tools your teams already use for decisions.

Related use cases

NL Querying works best alongside

Metric Attribution

When your NL query returns a metric movement, attribution automatically identifies which segments drove it.

Learn more

Stakeholder Reporting

Convert NL query results into formatted reports for executives and department leads without manual write-up.

Learn more

Let your team ask questions directly

Connect your warehouse with a read-only service account. Business teams get answers in minutes. Analysts get their time back.