Follow the workflow through discovery, planning, analysis, validation and delivery. Scroll through each step at your own pace.
Enable scroll
Insight Autopilot
From a question to a deployed, governed data product - run by 8 specialist agents
LiveActivity 4Save sessionSessionsAnnie
AIWhich customers are most likely to churn next quarter?
session: 5123af2d8b1e
AI
Ask a question to begin
Type a business question above and the agents will interpret it, discover the right data across your catalogs, plan the joins, profile the data, chart, and prepare a governed output - one step at a time.
AINot sure what to ask? Let's build a questionfrom your catalog
1 · Focus->2 · Question
What do you want to learn about?
Governance1
Cost1
Data quality1
Show all
What are the top causes of product returns by region?
Show monthly sales trends across channels
Which customers are most likely to churn next quarter?
AI
Agents at work
Done - review what the interpret, discovery, ontology and context agents found, then continue.
QWhich customers are most likely to churn next quarter?
100%
okInterpret AgentUnderstanding "Which customers are most likely to churn next quarter?"done
Parsed intent & time window
Identify customers at highest risk of discontinuing their relationship with the business in the next quarter.
Analyze historical churn patterns and behavioral indicators to build a predictive risk profile
Segment customers by churn probability score to prioritize retention efforts
Identify key drivers and early warning signals associated with customer churn
okDiscovery AgentScanning live catalogs & enginesdone
1 tables ranked across athena
mlforge_catalog.customer_churn
athena - 374 rows
okOntology AgentAligning to the operational ontologydone
okPlanning AgentStanding by to plan the join pathdone
Ready to plan the join path on demand
These results stay here until you continue - open the wizard to plan, profile, chart and build the dashboard
1 of 8 complete - 7 to go
IStep 1: InterpretUnderstand the question
Interpret Agent
Interpreted Intent
Identify customers at highest risk of discontinuing their relationship with the business in the next quarter
Sub-intents
Predict churn probability scores for each customer
Identify key behavioral or engagement signals that indicate churn risk
Segment customers by churn risk level (high, medium, low)
Determine the time frame and leading indicators specific to next quarter
Prioritize customers for retention intervention based on churn likelihood
AIAsk the agent to change this step
e.g. treat churn as 30-day inactivity
Chain of Thought
Parse Request
Reading the question: "Which customers are most likely to churn next quarter?". I read this as: Identify customers at highest risk of discontinuing their relationship with the business in the next quarter.
96%
Classify Framing -> Predictive
This is forward-looking, so the output should be a scored / projected result rather than a historical aggregate.
90%
Resolve Metrics & Grain
Pulled the measurable terms out of the question - metrics: churn rate; break-downs: customer. I'll look for these when I scan the catalog next.
84%
Decompose -> 5 Sub-intents
Splitting the question into: (1) Predict churn probability scores for each customer; (2) Identify key behavioral or engagement signals that indicate churn risk; (3) Segment customers by churn risk level (high, medium, low); (4) Determine the time frame and leading indicators specific to next quarter; (5) Prioritize customers for retention intervention based on churn likelihood.
88%
Agent Insights
*
Predictive, not descriptive
Identify customers at highest risk of discontinuing their relationship with the business in the next quarter
*
5 sub-intents
Predict churn probability scores for each customer; Identify key behavioral or engagement signals that indicate churn risk; Segment customers by churn risk level (high, medium, low); Determine the time frame and leading indicators specific to next quarter
Reporting and AI questions need more than prepared data. Ontology supplies business meaning; lineage supplies the path back to the source. Together, they give teams context for reviewing the result.
Three connected layers
01
Catalog mesh
Bring catalogs, BI, documents, and operational sources into a connected inventory with lineage across systems.
02
Meaning & trust
Align business terms and relationships, preserve where assertions came from, and keep disagreements visible for review.
03
Agentic workflow
Ground a question in shared context, confirm its scope, validate the evidence, and return an answer with gaps stated.
Product capability tour
Explore each capability working together.
Core intelligence and platform operations now live in one connected product collection.
Super Catalog
Live
Discovery
Assets
Lineage
Governance
Search assets across all catalogs
Search
HR payroll eventssnowflake
snowflake.public.payroll
Restricted
complaintsglue
auto_quality_db.complaints
Unknown
recallsglue
auto_quality_db.recalls
tier:gold
Super Catalog
Find trusted data across every catalog, understand where it came from and see who can use it.
Super Catalog
Live
Discovery
Assets
Lineage
Connectors
Registered types
glue
redshift
snowflake
unity
ConnectorTypeStatus
glue:931979956162glueenabled
unity:databricksunityenabled
redshift:kx-analyticsredshiftenabled
Platform & Ops 01
Catalog Federation
Unify Glue, Snowflake, Databricks and Redshift into one normalized catalog.
Knowledge Hub
Supported files
IngestionEntity Graph
478 entities · 975 docs
live · Neptune
Top entitiesMeridian · 12NovaTech · 12Dr. Kenji · 10
Structured + Unstructured Ontology
Connect tables, documents, meetings, files and operational records through shared business meaning.
Annie / Your AI assistant
Start with a question. Keep the business context.
Explore metrics, definitions, lineage and operational context in plain language with the AI assistant inside Datapane.
For analysts, business teams, administrators, developers and data scientists working with the same data foundation.
GenAI ETL supports data preparation. The ontology layer aligns business meaning, while lineage maps source dependencies. Report automation and Annie use that context to support reporting and investigation.
Which product should our team start with?
Start with the gap in your current workflow: preparing data, aligning definitions, tracing an output or delivering a report. Talk to our team to discuss your sources, users and requirements.
Can we explore the product before discussing a rollout?
The ontology model on this page is interactive and uses reference data. For a discussion of your own data, use Talk to Our Team; it opens our contact form.
Where can we see industry-specific applications?
Our Solutions pages describe workflows for manufacturing, banking, wealth, treasury and risk, with the original solution briefs available for direct download.
Your next step
Bring the workflow. We’ll discuss the fit.
Tell us what your team needs to prepare, understand or report. We’ll help identify the relevant products and implementation requirements.
