Breaks the question down
The engine plans the investigation the way an experienced analyst would, step by step.
Cognitive Depth Engine
For the questions that need more than one query.
Some questions cannot be answered with a single number. The Cognitive Depth Engine breaks a complex business question into steps, investigates each one across your data, and brings the findings together into one explained recommendation.
The problem
"Why did margin fall in the West?" is not one query. An analyst would check revenue, costs, freight, product mix and pricing, then connect the findings. Simple chat tools answer the first query and stop.
Investigates across domains
Each step can draw on different data, such as finance, operations and sales, so cross-functional causes are not missed.
Explains how it got there
The final answer shows every step, what each one found and how they lead to the conclusion.
Sample answer
What an answer looks like.
Illustrative example on sample data.
Which shipments are at risk of delay this week, and why?
Answer 3 of 3 models agree2 of 38 shipments are high risk, both with BlueLine, whose on-time rate fell to 71% this month. Re-booking SH-20417 protects a $420K customer order.
SQL it ran
SELECT s.shipment_id, s.lane, c.name AS carrier,
s.eta, r.delay_risk
FROM logistics.shipments s
JOIN logistics.carriers c ON c.id = s.carrier_id
JOIN ml.delay_risk_scores r ON r.shipment_id = s.shipment_id
WHERE s.status = 'IN_TRANSIT'
ORDER BY r.delay_risk DESC
LIMIT 10;Definitions used
- Delay risk = model score (0–1), refreshed hourly
Sources
logistics.shipmentsml.delay_risk_scoresTry asking
Questions teams ask.
“Why did West region margin fall in Q3?”
The drivers ranked by impact, such as a carrier change raising freight costs, with each step shown.
“What would happen to Q4 if freight rises 10%?”
The projected impact on margin by region, and the assumptions behind it.
“Which suppliers put our delivery targets at risk?”
The suppliers ranked by risk, the evidence for each, and suggested actions.
Connectors
Works with the data you already have.
Snowflake
BigQuery
Redshift
PostgreSQL
MySQL
SQL Server
Oracle
ClickHouse
Trino- Athena
SAP HANA & B1
Excel & CSV