AI services · Analytics

Free analysis on your data. Then ask it anything, in plain language.

Send us a data sample. We model your metrics, run a free analysis, and show you the answers you have been waiting days for. If it holds up, we connect your warehouse, build the natural-language layer, and set up the anomaly alerts and forecasts so the answers come to you.

Grounded in your real metrics Anomalies surfaced Forecasts with honest uncertainty
Sample analysis: 18 months of data
Metrics modeled 40
Anomalies surfaced 7
Forecast accuracy on backtest 92%
Time to an answer days to seconds
Reports automated weekly
Grounded in defined metrics Natural-language queries Anomaly alerts Forecasting Your warehouse
Why this exists

Most companies already have the data to answer their hardest questions. The wait is the problem.

The data is there. In BigQuery, in Snowflake, in Postgres tables that someone built two years ago. The problem is that getting an answer requires a data analyst, a ticket, a few days of back and forth, and a spreadsheet that nobody keeps current. By the time the answer arrives, the decision has already been made by instinct.

Decision intelligence changes the loop. You define your metrics once, with the business logic that makes them trustworthy. After that, anyone with access can ask a question in plain language and get the answer in seconds, grounded in the same numbers the data team would have pulled, with the anomalies surfaced and the forecast attached.

most
companies sit on the data to answer their hardest questions and still wait days for a person to pull it. The data warehouse is not the bottleneck. The bottleneck is the human in the middle.
How it works

Four steps. The free analysis comes back in days.

No platform to buy, no model to train from scratch, no six-month data strategy. We start with a sample you send us and come back with real answers on your data.

1

Connect a data sample

Share a sample of your warehouse data and the questions your team is asking today. Invoices, orders, web events, CRM data, anything with signal.

~1 day to set up
2

We model the metrics and backtest

We define the metrics layer: every KPI, its formula, and its grain. We run forecasts against your historical data and measure the error before you see anything.

A few days
3

Analysis review

You see the free analysis: real answers to your questions, anomalies flagged, forecasts with their accuracy measured on your own history. You keep it.

30 minutes
4

Roll out queries, alerts and briefs

If it holds up, we connect the warehouse, build the natural-language layer, wire the anomaly alerts, and schedule the automated exec briefs.

Scoped after the analysis
What the engagement covers

Ask in plain language. Catch anomalies early. Forecast what is next.

Three capabilities that compound. Each is useful on its own; together they make the data warehouse the fastest person in the room.

Ask in plain language

Type a question in English the way you would ask a colleague. The system translates it into SQL against your semantic metrics layer, runs it, and returns the answer with the chart. No dashboard to learn, no ticket to file.

Catch anomalies early

The system watches your key metrics continuously. When a number breaks its expected range, it fires an alert with the likely cause and the affected dimension, before anyone notices the revenue dip in the Monday review.

Forecast what is next

Demand, revenue, churn, capacity: forecasts run against your historical data, with honest confidence bands and the backtest accuracy shown upfront so you know what to trust and by how much.

Want the depth? How the semantic metrics layer prevents hallucinated numbers, how anomaly detection works, what makes a forecast trustworthy, and where this connects to your warehouse are all covered in the complete AI Analytics guide.

Sample query

This is what an answer looks like.

A plain-language question, the SQL it generated from your metrics layer, the result grounded in real numbers, and any anomaly the system detected along the way. Every number traces back to a defined metric, not a model guess.

The card on the right is illustrative. Your free analysis is built from your own data.

query · revenue_by_month
grounded in metrics layer anomaly flagged 2.1 s to answer
Question asked in plain language
"Why did revenue dip in March?"
SQL generated from your semantic metrics layer
SELECT month, SUM(net_revenue) AS revenue
FROM metrics.orders_monthly
WHERE month BETWEEN '2026-01-01' AND '2026-04-01'
  AND region = 'EMEA'
GROUP BY month ORDER BY month;
-- metric: net_revenue (defined: orders.amount - refunds.amount)
-- grain: monthly, region
Result + anomaly detected
March: EUR 184,200 vs Feb: EUR 241,800. Chart: line drops 24% in March. Anomaly: UK cohort refund rate spiked to 18% (baseline 4%). The model surfaced this dimension without being asked.
Where it pays off

The questions your team is asking a person to answer today.

If the answer lives in your warehouse and you are waiting more than a few minutes for it, that is the gap this fills.

NeedThe questionWhat it returns
Self-serve analyticsWhy did revenue dip in March?The answer and the chart, grounded in your defined metrics
Anomaly alertsTell me when something breaksAn alert with the likely cause and the affected dimension
ForecastingWhat will demand be next quarter?A forecast with a confidence band, backtested on your history
Exec briefSummarize the weekA written brief, generated from your metrics, sent on schedule
KPI watchAre we on track?A live read on every target, flagged when off course
Engagement tiers

From a free analysis to a running intelligence layer.

The analysis is always free. After that you pay for the build and the ongoing operation. The exact scope is set after the free analysis, because it depends on your warehouse and the metrics you want to define.

Pilot

Prove it

Free · on a data sample
  • Free analysis on a sample of your data
  • Metrics modeled and backtested
  • Anomalies surfaced on the sample
  • Answers to your top questions, grounded
Build

Connect the warehouse

Talk to us · fixed build
  • Everything in Pilot
  • Full warehouse connection and semantic metrics layer
  • Natural-language query interface for your team
  • Anomaly detection and alert routing
  • Automated weekly exec brief
Operate

Keep it honest

Talk to us · monthly
  • Everything in Build
  • Ongoing metric and forecast maintenance
  • Model retraining as your data changes
  • New query types and alert rules added
  • Monthly accuracy and adoption report
How pricing works. The Pilot is free and you keep the analysis. The Build is a fixed scope, agreed before any work begins. Operate is a monthly retainer that keeps the metrics and forecasts honest as your business and data change. No platform licence to buy.
Proof

We ship working analytics, not slideware.

AI Analytics is a data engineering and modeling engagement, which is the work we are built for. A few that map closely:

Request

Get a free analysis on your data.

Tell us your work email and where your data lives. We reply within 24 hours to set up the free analysis. You keep the result whether or not you continue.

  • Run on your own data, not a canned demo.
  • Real answers to your real questions, grounded in your metrics.
  • Anomalies and forecasts included in the analysis.
  • We build it, we run it. No platform to license.

Prefer a longer brief? Use our form

Get a free data analysis

Two fields. We do the rest.

No credit card. No commitment. We reply within 24 hours with next steps.

FAQ

Questions we get every week.

Will the AI invent numbers that are not in my data? +
No. Every answer is grounded in a semantic metrics layer built from your own definitions. The system generates SQL against your real tables and returns only what the query finds. If a number is not in your data, it does not appear in the answer. This is the fundamental difference between a chatbot on a raw database and a decision intelligence system.
Do I need a data team to use this? +
No. We build the metrics layer and connect it to your warehouse as part of the engagement. After that, business users ask questions in plain language and get answers. The technical scaffolding is ours to build and maintain.
Which databases and warehouses does it connect to? +
BigQuery, Snowflake, and Postgres are the primary targets. We can extend to Redshift, Databricks, and other ANSI-SQL-compatible systems. The free analysis runs on a sample you send us, so connectivity is confirmed before any production work begins.
How accurate are the forecasts? +
We backtest every forecast model on historical data before shipping it, so you see the measured error on your own data, not an industry benchmark. Forecasts include honest confidence bands rather than single-point predictions. Where uncertainty is high, the system says so.
Who can see which data and queries? +
Access control is built into the metrics layer. You define which teams see which metrics, and the query layer enforces that at runtime. No one can ask a question that returns data they are not permitted to see.
What does it cost? +
The Pilot tier is free and runs on a data sample you provide. The Build and Operate tiers are priced after the free analysis, because the scope depends on your warehouse, the number of metrics you want to define, and how many users will be running queries. We quote a fixed number before any work begins.

Ask your data the questions you have been waiting days to answer.

A free analysis on your real data, the metrics modeled and backtested, the anomalies surfaced. Then we build the layer that makes every answer take seconds.

Get a free analysis
Free analysis · You keep the result · 24h reply

Related reading

Related reading

Get a free analysis