Text to SQL

Ask the data. Get the agreed number.

Plain questions are answered from your approved metrics, never from a guess, with the query shown.

About 5 weeks
to live on real traffic
Founder-led
no handoff to juniors
Fixed price
agreed before work starts
RORevenue by region last quarter, with the number of customers
Query written from your question Read only, 1 table you can see
SELECT region, SUM(amount) AS revenue, COUNT(DISTINCT customer_id) AS customersFROM sales.ordersWHERE order_date >= DATE '2026-07-01' AND order_date < DATE '2026-10-01'GROUP BY regionORDER BY revenue DESC;
Result 4 rows in 0.8sSave to dashboard
Running The result shows here once the query passes its checks.
RegionRevenueCustomers
North America$4,812,3001,204
Europe$3,206,950988
Asia Pacific$1,944,120612
Latin America$702,480231
North America
Europe
Asia Pacific
Latin America
The problem

Every simple question waits days in the analyst queue

When people try a chat tool instead, it makes up its own definition of revenue, and two answers to the same question disagree.

67%

of decision makers do not fully trust the data they rely on.

Sourceprecisely.com

54%

of business leaders are not sure the data they need is accessible.

Sourcesalesforce.com

26%

of their data is untrustworthy, say data and analytics leaders.

Sourcesalesforce.com

How we fix it

Anyone can ask the data and get the number finance agrees with.

What changesWithout itWith it
WaitingDays in the analyst queue.An answer in a minute.
DefinitionsEach dashboard has its own.One agreed metric for everyone.
TrustNo way to check the number.The query shown with every answer.
How it works

What we build

What changes for your teams. We pick the tools to fit the systems you already run.

Query written from your question Read only, 1 table you can see
SELECT region, SUM(amount) AS revenue, COUNT(DISTINCT customer_id) AS customersFROM sales.ordersWHERE order_date >= DATE '2026-07-01' AND order_date < DATE '2026-10-01'GROUP BY regionORDER BY revenue DESC;
What changes

Measured against where you started

Example: Acme Inc

What changes at a company like Acme Inc

Illustrative figures, not a client result. Your own baseline is taken in the first week.

MeasureBeforeAfter
Time to answer a routine question3 daysUnder 1 minute
Routine questions in the analyst queueMost of the queueA few
Answers that match the agreed metricUnknownEvery tested question
Questions in the test setNoneAbout 150

How others built it

Another company built the same kind of system and wrote publicly about how it went.

What you keep

Everything runs in your cloud, and stays yours

You own the config, the tests and the runbooks, so you can swap any part later.

The question layer running in your cloud, with a read-only role.

Result 4 rows in 0.8sSave to dashboard
Running The result shows here once the query passes its checks.
RegionRevenueCustomers
North America$4,812,3001,204
Europe$3,206,950988
Asia Pacific$1,944,120612
Latin America$702,480231
North America
Europe
Asia Pacific
Latin America

Agreed metric definitions in code, owned by your data team.

Query written from your question Read only, 1 table you can see
SELECT region, SUM(amount) AS revenue, COUNT(DISTINCT customer_id) AS customersFROM sales.ordersWHERE order_date >= DATE '2026-07-01' AND order_date < DATE '2026-10-01'GROUP BY regionORDER BY revenue DESC;
  • A test set of real questions that runs on every change.

  • A runbook your data team owns.

Typical building blocks: dbt Semantic Layer and dbt MCP, or Cube; an agentic SQL fallback with schema pruning, dry run and self-repair; a read-only role; a golden question set in CI.

Alternatives: WrenAI; Snowflake Cortex Analyst; Databricks Genie (Vanna is archived).

How the weeks run

Live in about 5 weeks

The work happens in your cloud, with the people who will own it.

  1. Week 1Collect real questions and agree the metric definitions with your data team.
  2. Weeks 2 to 4Build the question layer on your warehouse and tune it against the test questions.
  3. Week 5Open it to the first teams and hand over the runbook.
  4. After thatRunning on real traffic

    Plain questions answered from agreed metrics.

Before we start

  • A read-only role on the warehouse.
  • A named data owner who agrees each metric definition.
  • About 4 hours a week from your data team.
Fit

Is this the right first build?

It fits when

  • Every question waits for an analyst.
  • Two dashboards give two answers.
  • A team on your side will own it after we leave.

It does not fit when

  • You want a hosted tool picked and resold to you.
  • You want a strategy deck in place of working software.
  • Nobody on your side can own it after we leave.

What’s included and what isn’t

Included

  • One warehouse.
  • Up to 25 agreed metric definitions.
  • Up to 3 subject areas, such as sales, finance and product.
  • A test set of up to 100 real questions with known answers.

Not included

  • Building or cleaning the warehouse itself.
  • Dashboards and scheduled reports.
  • Write access of any kind.
Fixed quoteOne price, agreed before work starts

About 5 weeks. You get a price range before the call and one fixed number after it. That number is the invoice.

  • If the work takes longer than we planned, we absorb the cost.
  • You pay 50% at the start and 50% at acceptance. Builds of 6 weeks or more pay 40% at the start, 30% at the midpoint and 30% at acceptance.
  • If a build starts within 60 days of your Audit results, the full Audit fee comes off its price.
  • We start three new projects a quarter, so your start date depends on when you book.
Book a demoSee starting prices
FAQ

Questions teams ask before they book

Which tools do you use?

We pick the one that fits the systems you already run. You own the config, the tests and the runbooks, so you can swap any part later. Most often the dbt Semantic Layer with dbt MCP, or Cube; WrenAI, Snowflake Cortex Analyst or Databricks Genie when they fit better.

Can it change our data?

No. It runs with a read-only role, and each query is checked with a dry run before it runs.

What if a question has no agreed metric?

It says so, writes a draft query and sends it to an analyst. It does not guess a definition.

Which warehouses does it work with?

The common cloud warehouses, through the semantic layer you already have or one we set up with you.

How do you know the answers are right?

Your data team writes real questions with known answers. Every change must pass them.

Text to SQL, running in your cloud in about 5 weeks.

Plain questions answered from agreed metrics. You own the config, the tests and the runbooks.