How It Works
From Raw Data to
Causal Intelligence
Three steps. Connect your data warehouse, build a causal knowledge graph, and analyze with plain language. Fig handles the complexity so you focus on decisions.
Connect
Plug In Your Data Warehouse
Connect Snowflake, BigQuery, or PostgreSQL with read-only credentials. Fig scans your entire schema, discovers tables, and profiles your data — column types, null rates, value distributions, and join paths. No data leaves your warehouse.
Schema scanning
Every database, schema, and table discovered automatically.
Table profiling
Row counts, null rates, and value distributions at a glance.
Column type detection
Metrics, dimensions, dates, and IDs classified by AI.
Relationship discovery
Foreign keys and join paths identified across tables.
Build Knowledge Graph
Map Causal Relationships
Fig's KG Builder agent constructs a knowledge graph that connects metrics to dimensions to tables to columns. It maps causal relationships — Revenue depends on Customer Count depends on Retention Rate — and applies industry templates as a starting point. You review every node and edge before it goes live.
Human-in-the-loop. Every metric, dimension, and relationship the AI proposes must be approved by your team before it enters the graph.
Automatic node discovery
Metrics, dimensions, KPIs, and tables detected from your data.
Causal relationship mapping
Revenue depends on Customer Count depends on Retention — all mapped.
Industry templates
Pre-built ontologies for healthcare, retail, finance, and more.
Human-in-the-loop review
Every proposed node and relationship is reviewed before it's live.
Analyze
Ask Anything in Plain Language
Ask “Why did revenue drop last quarter?” and Fig selects the right analysis algorithm, traces causal chains through the knowledge graph, and delivers an answer with evidence — data tables, charts, SQL lineage, confidence levels, and actionable recommendations.
Example Prompt
“Why did customer churn increase 15% last month, and which segments are most affected?”
Natural language queries
Ask 'Why did revenue drop?' and get a structured answer.
Algorithm auto-selection
Fig picks the right algorithm from 12 built-in options.
Evidence-backed answers
Every conclusion comes with data, SQL lineage, and confidence levels.
Structured reports
Charts, tables, causal chains, and actionable recommendations.
What Makes Fig Different
Not another dashboard tool. Fig adds the causal intelligence layer your data stack is missing.
Causal, Not Correlational
Fig maps how metrics actually cause and affect each other — not just which ones move together. When revenue drops, you see the full causal chain, not a scatter plot.
12 Built-In Algorithms
Root cause analysis, anomaly detection, concentration analysis, scenario planning, and 8 more. Fig selects the right algorithm automatically based on your question.
Knowledge Graph Foundation
Every analysis is grounded in a knowledge graph that maps your specific business structure — not generic patterns. Context-aware from day one.
Evidence, Not Guesses
Every answer includes the data it's based on, the SQL that produced it, confidence levels, and the causal chain it traced. You can verify everything.
See It in Action
Connect your data warehouse and go from raw data to causal intelligence in under 5 minutes.