Fabric
AI-native data fabric that transforms fragmented sources into clean, connected, analysis-ready streams.
At a glance
Connect any source
Databases, object storage, APIs, files, and web data.
Auto-clean and unify
AI resolves schema drift, duplicates, and messy text.
Analysis-ready output
Consistent tables, views, and events with lineage.
Operates in your cloud
Security, privacy, and performance under your control.
Works with the suite
Guard for governance, Studio for GenAI, Signals for decisions.
Why Fabric
Enterprises live with scattered systems, inconsistent formats, and slow data onboarding. Fabric turns this complexity into clarity. It reduces time spent on integration and quality work, creates a unified layer for analytics and GenAI, and establishes trust in the numbers.
What Fabric does
Connect
Native connectors for databases, storage, and APIs with secure credential management.
- •Native connectors for PostgreSQL, MySQL, MSSQL, MongoDB, S3-compatible storage, Google Cloud Storage, Azure Blob, FTP/SFTP, REST and GraphQL APIs, CSV and Excel.
- •Secure credentials vault, secrets rotation, and scoped access.
- •Incremental sync, change data capture where supported.
Clean
AI-assisted profiling and automatic data quality improvements.
- •AI-assisted profiling detects anomalies, outliers, and missing values.
- •Automatic normalization, type inference, unit unification, and text standardization.
- •Deduplication and entity resolution with learned matching rules.
Model
Schema unification and semantic layer for consistent analytics.
- •Schema unification to a documented "golden" model.
- •Semantic layer definitions for metrics, dimensions, and time grains.
- •Reusable transformations and templates for common domains.
Deliver
Analysis-ready tables and streams with monitoring and versioning.
- •Analysis-ready tables and streams to your warehouse or lake.
- •Real-time or batch delivery with SLAs and monitoring.
- •Versioned datasets with rollback and reproducibility.
AI inside Fabric
Structure from chaos
LLM-powered adapters infer schemas, map fields, and generate transformation templates.
Quality by learning
The system learns from human fixes then applies them at scale.
Natural language assist
Describe a table you need, Fabric proposes the pipeline and the contract.
Explainability
Every change includes reasons, confidence, and source references.
Architecture overview
Three-layer architecture ensuring scalability, security, and observability.
Three-layer architecture that runs in your VPC. Policies enforced by Guard. Lineage and SLAs observed end to end. Studio and Signals sit on top for knowledge and action.
Common use cases
Finance operations
Consolidate ledger, billing, and banking feeds into a single model for daily closes.
RevOps
Unify CRM, marketing, and product events for pipeline, churn, and attribution.
Supply and inventory
Normalize vendor and logistics data for fulfillment and risk alerts.
People analytics
Merge ATS, HRIS, and payroll for hiring and retention insights.
Public data enrichment
Ingest web and third-party data to augment internal truth.
Example templates
Finance ledger unification
Map ERP, billing, and bank feeds to a common chart of accounts.
B2B funnel model
Leads to revenue with multi-touch attribution.
Events normalization
Product telemetry to a clean session-event schema.
Vendor inventory
SKU merging and stock normalization across suppliers.