Analytical stores
Snowflake, Databricks, Microsoft Fabric, BigQuery, Redshift, Synapse, Postgres, SQL Server.
CONNECTORS & DATA FOUNDATION
Meet the systems where they are, including the ones from 1994.
Enterprise data lives in a warehouse, a mainframe, six SaaS products, a SharePoint site and someone's quarterly export. Flexday connects to all of it and turns it into something an agent can reason over safely.
COVERAGE
Where the data comes from.
Connectors carry structure and permissions, not just rows. What a user can see in the source is what they can see through Flexday.
WAREHOUSE & LAKE
SYSTEMS OF RECORD
ERP & core business
SAP, Oracle, Dynamics, NetSuite, Salesforce, Workday, ServiceNow, Coupa, Ariba.
LEGACY & OPERATIONAL
The awkward ones
AS/400 and IBM i, mainframe extracts, EDI feeds, imaging systems, handheld and telematics feeds, SCADA and historians.
CONTENT
Documents & media
SharePoint, OneDrive, Google Drive, S3, Blob, Confluence, network shares, email archives, video libraries.
EXTERNAL
Market & third-party
Syndicated retail measures, supplier risk feeds, credit and firmographic data, weather, commodity and freight indices, public web.
DELIVERY
Where answers land
Teams, Slack, Outlook, Power BI, Tableau, your own applications via REST, webhooks and embeddable components.
NO CONNECTOR?
Build one
A custom connector is a full-code object like any other, schema, auth, permissions, profiling. Usually a day or two, not a procurement cycle.
NO ACCESS YET?
Start anyway
Documents, exports and synthetic sets get a real solution built while your access request works its way through. Swap the source in later.
FACT BASES
Curated corpora, not a document dump
A fact base is a governed body of knowledge for a specific purpose: this policy set, these contracts, this archive. It carries its own chunking and citation policy, its own permissions, its own freshness rules and its own evaluation set.
- Entitlements inherited from the source system, enforced at query time
- Freshness and staleness tracking with alerting on drift
- Versioned, you can ask what the system would have answered last quarter
- Reusable across solutions without duplicating the content
DATA MODELS
Semantics that outlive the solution
Data models bind ontology concepts to physical sources. Because agents query the model rather than the table, you can re-platform underneath a live solution and nothing above it needs to change.
- Ontology-bound: a concept maps to many sources, and sources can be swapped
- Measures defined once, in Layer 3, so every solution reconciles
- Lineage from any figure back through the model to the originating row
- Exportable to dbt, Unity Catalog and Snowflake semantic definitions
DATA FOUNDATION SPECIFICATIONFLEXDAY.AI
- INGESTION
- Batch, incremental, streaming, event-driven, and on-demand federated query
- PERMISSIONS
- Source-system entitlements propagated to query time; no post-hoc filtering of results
- PROFILING
- Automatic data portraits with schema, semantic, quality, freshness and join inference
- LINEAGE
- Row-level provenance from answer to source, retained for audit
- CUSTOM CONNECTORS
- Full-code connector SDK with auth, schema, permission and profiling contracts
- WITHOUT ACCESS
- Document-, export- and synthetic-first builds with a mapped path to live sources
Your data isn't ready. Start anyway.
We build on documents, exports and synthetic sets, then swap in live sources without a rewrite.