# Anjana Data Governance

> Data and AI governance platform with a configurable metamodel, catalog, business glossary, BPMN approval workflows, and a Marketplace where Data Sharing Agreements grant access in Snowflake, Databricks, Power BI and other platforms; self-hosted, SaaS or hybrid.

Source: https://toolsfordata.com/tools/anjana-data/ · Updated 2026-11-19

[Website](https://anjanadata.com/)

Best for: Spanish and European enterprises and public bodies that want a governance layer independent of any one platform, with automated access provisioning, AI Act templates, and self-hosted or EU-hosted deployment.

## Overview

- Core Strength: Approved data sharing agreements turn into real grants in Snowflake, Power BI and cloud IAM — plus AI Act and ISO 42001 templates
- Lineage Depth: Partial — table, column, end-to-end, impact analysis
- Catalog Curation: Partial — glossary, certification, custom fields, propagation
- Stewardship Workflows: Partial — review/approval, campaigns
- Policy & Classification: Partial — policy, enforcement
- Compliance Depth: Partial — audit trail, regulatory mapping, RoPA
- Marketplace & Data Products: Full — data products, access requests, data mesh, subscriptions
- Built-in Data Quality: Partial — contracts
- AI & Agent Interface: Partial — MCP, knowledge graph, documents
- AI Governance: Partial — AI inventory, AI lineage, AI regulation
- Developer Surface: Partial — API
- Deployment Model: Full — SaaS, self-hosted, hybrid, data residency
- Security & Access: Partial — SSO, RBAC, audit log, certifications
- Notable Caveat: No built-in AI, and lineage is documented by hand rather than parsed from SQL or dbt

## What it's used for

Anjana Data is a Spanish data and AI governance platform that sits on top of existing data platforms. Teams use it to run a governance operating model, a catalog and glossary, and a Marketplace where approved data sharing agreements grant access.

## Marketplace

An approved Data Sharing Agreement creates real grants in Snowflake, Denodo, Power BI, S3, Azure Storage, Ranger and cloud IAM, and revokes them when it expires.

## Stewardship

A configurable metamodel, domains, roles and BPMN workflows run every create, change, deprecate and access step, with versions and expiry dates on each asset.

## AI governance

Templates model AI systems, use cases, risk and impact assessments against the EU AI Act and ISO 42001, and link models to their training datasets.

## Catalog

Plugins extract metadata from Oracle, SQL Server, Snowflake, Unity Catalog, BigQuery, Redshift, Glue, Denodo, Power BI and Tableau.

## AI

Anjana ships no AI of its own. Its MCP server (version 26.1, June 2026) lets your AI client search, enrich and request access with your permissions, with all changes left as drafts.

## Pricing

Anjana prices by governed assets, with no per-user fees. On AWS Marketplace, SaaS starts at $37,800 a year.

## Deployment

Anjana runs self-hosted on VMs or Kubernetes, as SaaS on AWS, Azure or GCP in EU or US regions, or hybrid.

## Security

Anjana supports SAML and OIDC SSO, permissions by domain and an audit module, with Spain's ENS certificate. It has no SOC 2 or ISO 27001 of its own.

## Integrations

Anjana Data connects to Snowflake, Databricks, BigQuery, Redshift, Power BI and Tableau, and to Apache Ranger to push access policies.

## Review

Strengths:
- Approved data sharing agreements create grants in Snowflake, Power BI, S3 and cloud IAM, and expiry revokes them.
- Self-hosted, SaaS or hybrid, priced by governed assets with no per-user fees.
- AI Act and ISO 42001 templates, plus RoPA and DPIA asset types.

Caveats:
- No built-in AI: no chat, generated descriptions or agents.
- Lineage is documented by hand, not parsed from SQL, dbt or OpenLineage.
- No automated PII scanning, profiling or data quality checks.

Recommendation: Choose Anjana for a governance operating model with access provisioning and AI Act templates, especially in Spain. For automated lineage and AI features, consider Collibra or Atlan.

## Facts

- Anjana Data was founded in Bilbao in 2019.
- Telefónica Tech started reselling Anjana in Spain in April 2025.
- Anjana 26.1 (June 2026) added a new Marketplace, an MCP server and a Kubernetes install.

## Pricing

- Model: per governed asset
- Entry team price: $37,800/year
- Entry enterprise price: Custom
- Source: https://aws.amazon.com/marketplace/pp/prodview-fizkdh2ctxv6o (checked 2026-11-12)

## Features

### Data Catalog

- Automated Metadata Harvesting: Yes — Tot plugins discover and extract tables, columns, views and reports from Oracle, SQL Server, Snowflake, Databricks Unity Catalog, BigQuery, Redshift, AWS Glue, S3, Denodo, Power BI and Tableau; a user picks which structures to import in an extraction wizard. ([source](https://wiki.anjanadata.com/en/funcional/26.1/metadato-automatico))
- Search & Discovery: Yes — The Catalog search bar runs across all attributes and relationships, with static, dynamic and quick filters by subtype, status, organizational unit and adherence, plus saved searches. ([source](https://wiki.anjanadata.com/en/funcional/26.1/anjana-portal))
- AI-Powered Search: No — Search runs on Solr with attribute weights, stopwords and a synonyms file; there is no built-in natural language search, and plain-English queries only work through an external AI client connected to the MCP server. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/search-configuration))
- Usage & Popularity Statistics: No — The docs show no per-asset query counts, view counts or popularity ranking in the Catalog. ([source](https://wiki.anjanadata.com/en/funcional/26.1/anjana-portal))
- Custom Fields & Templates: Yes — Administrators define metadata attributes (text, select, date, entity reference and more) and assign them to templates per entity subtype, with sections, menus and validations. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/atributos-de-metadatos))
- Asset Certification: Yes — Each asset moves through lifecycle statuses (draft, pending validation, approved, deprecated, expired) set by validation workflows, and the status icon is shown on every asset. ([source](https://wiki.anjanadata.com/en/funcional/26.1/acciones-sobre-los-objetos))
- Business Glossary: Yes — Business terms, metrics and other glossary entities are owned by data domains, approved through validation workflows and linked to catalog assets through configurable relationships. ([source](https://wiki.anjanadata.com/en/funcional/26.1/glossary-of-anjana-data-terms))
- Metadata Propagation: Yes — HERITABLE attributes pass boolean values between related objects, for example a dataset is flagged as holding personal data when any of its fields is, and a DSA inherits values from the datasets it contains. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/validaciones-sobre-plantillas))

### Data Lineage

- Table-Level Lineage: Yes — Datasets are linked as inputs and outputs of processes and process instances, which shows which tables feed which downstream datasets. ([source](https://wiki.anjanadata.com/en/funcional/26.1/linaje))
- Column-Level Lineage: Yes — Dataset fields can be related to other fields, metrics and reports, which shows which fields are used to build a report; these links are relationships in the metamodel, not parsed from SQL. ([source](https://wiki.anjanadata.com/en/funcional/26.1/dataset_field))
- End-to-End Lineage: Yes — Relationships chain source datasets through processes and solutions to reports and business terms, so the graph can run from source system to report, though it is built from documented relationships rather than query parsing. ([source](https://wiki.anjanadata.com/en/funcional/26.1/conceptos-basicos))
- Impact Analysis: Yes — Versioning rules and lineage are used to review the impact of a change before it is approved, and deprecating an asset notifies the users who rely on it. ([source](https://wiki.anjanadata.com/en/funcional/26.1/conceptos-basicos))
- Automated Lineage Extraction: No — Lineage is drawn from relationships between entities, processes and reports in the metamodel; the plugins discover tables, fields, views and reports, but none parse SQL, query logs or ETL jobs to build lineage. ([source](https://wiki.anjanadata.com/en/funcional/26.1/linaje))
- OpenLineage Support: No — No OpenLineage integration appears in the plugin catalog or the docs. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/tot-plugins))

### Data Stewardship

- Critical Data Elements: No — There is no dedicated critical data element register; a team would have to model it as a custom attribute or entity type. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/metamodelo-y-plantillas-de-metadatos))
- Review & Approval Workflows: Yes — Validation workflows written in BPMN XML route creation, modification, deprecation, organizational unit changes and access requests to the right approvers, with comments and request reasons. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/workflows-flujos-de-validacion))
- Governance Campaigns: Yes — The Workspace and bulk attribute editing apply changes, assign owners or launch processes on many assets at once, and Excel editing updates a whole subtype in one upload. ([source](https://wiki.anjanadata.com/en/funcional/26.1/edicion-masiva-de-atributos))
- Governance Scorecards: No — No questionnaire-driven maturity scoring of assets or domains is documented. ([source](https://wiki.anjanadata.com/en/funcional/26.1/guia-de-usuario))
- Metadata Automation: Yes — Workflow listeners fire on governance events, an inventory interceptor transforms attribute values on save, and scheduled batches handle expirations and notifications. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/platform-notifications-and-workflows))

### Data Classification

- Sensitive Data Discovery: No — No automated PII scanning or pattern-based classifier is documented; personal-data flags on dataset fields are filled in manually or imported from source tags. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/tripletas-para-plugins))
- Data Policy: Yes — Access to governed data goes through Data Sharing Agreements: a consumer adheres to a DSA, the request is approved by workflow, and role and domain permissions control what each user can see and do. ([source](https://wiki.anjanadata.com/en/funcional/26.1/dsa))
- Native Policy Enforcement: Yes — When a DSA is approved, plugins create roles or groups and grant or revoke access in Snowflake, Denodo, Power BI, S3, Azure Storage, Ranger and cloud IAM, then revoke it when the DSA expires. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/snowflake-native-api))
- Data Masking: No — Plugins only obfuscate columns flagged as personal data in the limited data samples shown in the portal; there is no dynamic or static masking of query results. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/tripletas-para-plugins))

### AI Features

- AI Chat Assistant: No — There is no built-in chat; users bring their own AI client (for example Claude or GitHub Copilot) and connect it to the MCP server. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/mcp))
- AI Documentation Generation: No — The platform has no LLM of its own to draft descriptions; skills for AI clients to prepare upload files are listed as coming soon. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/ai))
- Agentic Metadata Enrichment: No — No agents are provided; an external AI client can create and enrich assets through the MCP, but every change stays in draft until a person approves it. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/mcp))

### Data Compliance

- Audit Trail: Yes — The traceability log records who created, modified, approved or requested access to each asset, with workflow comments and a downloadable snapshot of the asset after each action. ([source](https://wiki.anjanadata.com/en/funcional/26.1/auditoria-de-la-aplicacion))
- Compliance Reporting: Unknown — The website describes governance dashboards and reports through external BI tools and the AI governance setup stores conformity evidence, but no regulator-ready report or evidence pack is documented. ([source](https://anjanadata.com/funcionalidades/))
- Regulatory Framework Mapping: Yes — A preconfigured AI management system maps AI systems, risks, controls and evidence to the EU AI Act and ISO 42001 Annex A, and links GDPR processing activities to DPIAs and breaches. ([source](https://anjanadata.com/de-los-pilotos-de-ia-al-cumplimiento-del-ai-act-una-aproximacion-pragmatica-al-gobierno-de-la-ia/))
- Data Activity Monitoring: No — The audit module can store access events that governed systems push through the API, but Anjana does not collect or analyze source query activity itself. ([source](https://wiki.anjanadata.com/en/funcional/26.1/auditoria-de-la-aplicacion))
- ROPA: Yes — Records of Processing Activities are a regulatory compliance asset type in the Catalog, linked to DPIAs and data assets. ([source](https://wiki.anjanadata.com/en/funcional/26.1/anjana-portal))
- Data Retention Policies: No — Assets and DSAs have expiration dates that revoke access, but no rule archives or deletes the underlying data. ([source](https://wiki.anjanadata.com/en/funcional/26.1/acciones-sobre-los-objetos))
- Data Subject Rights: No — No DSAR, deletion or portability request handling is documented. ([source](https://wiki.anjanadata.com/en/funcional/26.1/guia-de-usuario))

### Collaboration

- Alert Notifications: Yes — The Hermes service sends governance notifications (workflows, expirations, deprecations, adherence failures) to Slack, Microsoft Teams, Google Chat and email. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/notification-providers))
- Discussions & Annotations: No — Comments exist only inside validation workflow steps; there are no threaded discussions on assets. ([source](https://wiki.anjanadata.com/en/funcional/26.1/workflows-submodules-and-validation))
- Ticketing Integrations: Unknown — The integrations page shows Jira, Confluence and ServiceNow logos, but the docs describe no plugin that creates tickets from governance events. ([source](https://anjanadata.com/arquitectura-e-integraciones/))

### Data Marketplace

- Self-Service Access Requests: Yes — Consumers add assets to a cart and request access; the adherence request routes to the DSA owner by workflow and, once approved, a plugin grants access in the source platform. ([source](https://wiki.anjanadata.com/en/funcional/26.1/marketplace-discovery-and-access-request))
- Data Products: Yes — Data Sharing Agreements package datasets, reports, KPIs and quality rules under one owner, with a contract, expiration date and versions, and are published in the Marketplace. ([source](https://wiki.anjanadata.com/en/funcional/26.1/dsa))
- Data Mesh Support: Yes — Data domains (organizational units) own their assets and workflows under a centralized, federated or hybrid governance model, and the vendor positions the Marketplace for data mesh setups. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/data-domains-organizational-units))
- Subscription Management: Yes — Users who adhere to a dataset or DSA are tracked as consumers and are notified when it is going to be deprecated or expire. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/platform-notifications-and-workflows))

### Data Quality

- SQL Table Tests: No — Quality rules are defined in Anjana but executed by external engines such as the Aqtiva (Omma Data) plugin, which is now deprecated. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/omma-data-aqtiva))
- Data Health Scores: No — No composite quality or health score per asset or domain is documented. ([source](https://wiki.anjanadata.com/en/funcional/26.1/guia-de-usuario))
- Anomaly Detection: No — No statistical or ML anomaly detection on data values or volumes is documented. ([source](https://wiki.anjanadata.com/en/funcional/26.1/guia-de-usuario))
- Data Profiling: No — Plugins return a limited data sample for preview but compute no column statistics such as null rates or distributions. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/tripletas-para-plugins))
- Data Contracts: Yes — DSA contracts are producer-consumer agreements with license terms, usage, quality and availability conditions; consumers must sign the contract before access is granted. ([source](https://wiki.anjanadata.com/en/funcional/26.1/glossary-of-anjana-data-terms))

### AI Governance

- AI Asset Inventory: Yes — The metamodel includes AI models and AI systems as entity types with their own templates, owners, required attributes and workflows. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/metamodelo-y-plantillas-de-metadatos))
- AI Lineage: Yes — The AI governance setup links each model to its training, validation and production datasets through relationships shown in the lineage graph. ([source](https://anjanadata.com/de-los-pilotos-de-ia-al-cumplimiento-del-ai-act-una-aproximacion-pragmatica-al-gobierno-de-la-ia/))
- AI Regulatory Compliance: Yes — Preloaded vocabularies follow the AI Act Article 6 risk classes, with entity types for risk assessments, impact assessments, GPAI training content summaries and conformity files. ([source](https://anjanadata.com/de-los-pilotos-de-ia-al-cumplimiento-del-ai-act-una-aproximacion-pragmatica-al-gobierno-de-la-ia/))
- AI Model Health Score: No — No composite trust or readiness score for AI models is documented. ([source](https://anjanadata.com/de-los-pilotos-de-ia-al-cumplimiento-del-ai-act-una-aproximacion-pragmatica-al-gobierno-de-la-ia/))
- AI Agent Access Governance: No — The MCP server applies the user's own permissions and keeps agent changes in draft, but it serves only metadata and adds no policy check at the point an agent reads data. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/mcp))

### Master Data Management

- Master Data Management: No — Anjana is a governance layer over existing platforms and does not match, merge or master records. ([source](https://anjanadata.com/anjana-data-platform/))

### Workflows

- REST / GraphQL API: Yes — Public, administrative and configuration REST APIs, documented in Swagger, read and write assets, relationships, workflows, notifications and audit entries. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/api))
- dbt Integration: No — There is no dbt plugin in the plugin catalog. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/tot-plugins))
- Python SDK: No — The SDK is a Java 17 / Spring Boot kit for building plugins; there is no pip-installable library. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/sdk))
- Webhooks: No — Outbound webhooks only deliver notifications to Slack, Microsoft Teams and Google Chat; there is no generic webhook to arbitrary endpoints. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/notification-providers))

### Data Context Layer

- MCP Server: Yes — The MCP server, first released in version 26.1 (June 2026), lets AI clients search assets, read lineage, catalog and enrich metadata, request DSA access and query the audit log with the user's own permissions. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/mcp))
- Knowledge Graph: Yes — Metadata is modeled as a graph of entities, users and organizational units linked by typed relationships, which the Lineage module explores from any node. ([source](https://wiki.anjanadata.com/en/funcional/26.1/linaje))
- Semantic Models: No — The glossary links business terms to assets, but Anjana does not generate a machine-readable semantic model with join paths. ([source](https://wiki.anjanadata.com/en/funcional/26.1/glossary-of-anjana-data-terms))
- Verified Queries: No — There is no built-in assistant; the MCP server exposes search, lineage, cataloging and access-request operations to an external AI client, with no store of approved example queries. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/mcp))
- Semantic Layer Interop: No — The Power BI plugin maps semantic models only to manage access permissions; no plugin reads or writes metric definitions from dbt, Snowflake, Databricks, Cube, LookML or OSI. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/powerbi))
- Unstructured Data Catalog: Yes — The Databricks plugin harvests Unity Catalog volumes as a separate "unstructured data asset" type (DBK_VOLUME) that can be governed like other assets, but only at volume level; the file plugins handle structured formats (CSV, Excel, Avro, Parquet), and documents, SharePoint and vector indexes are not covered. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/databricks))

### Security

- SSO: Yes — Login supports SAML 2.0 and OpenID Connect, with guides for Entra ID, Okta, Auth0, AWS, GCP, Keycloak and LDAP/Active Directory. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/sso-saml-2-0))
- RBAC: Yes — Roles hold permissions per entity type and action (create, modify, change status, deprecate, change organizational unit) and can be scoped by data domain. ([source](https://wiki.anjanadata.com/en/configuracion/26.1/permissions))
- Audit Log: Yes — The Audit module logs every user action in the platform, including changes made by AI clients through the MCP, and is restricted to roles with audit permission. ([source](https://wiki.anjanadata.com/en/funcional/26.1/auditoria-de-la-aplicacion))
- Compliance Certifications: Yes — Anjana Data S.L. holds an ENS (Esquema Nacional de Seguridad) medium-level certificate for 2026; for ISO 27001 and SOC 2 its trust center points to AWS, Google and Atlassian, not its own audits. ([source](https://anjanadata.com/wp-content/uploads/2026/08/CERTIFICADO-ENS-ANJANA-DATA-S.L.-MEDIA-2026_signed.pdf))
- SCIM Provisioning: No — Users are synced on a schedule by reading directory APIs (Microsoft Graph, Google Directory, Okta, LDAP/AD, Keycloak, AWS), not through a SCIM 2.0 endpoint. ([source](https://wiki.anjanadata.com/en/integraciones/26.1/user-provisioning))

### Deployment

- SaaS Deployment: Yes — The SaaS option is fully run by Anjana Data on AWS, Azure or GCP, with tiered capacity packages and a dedicated Customer Success Manager. ([source](https://wiki.anjanadata.com/en/solucion-tecnica/26.1/saas-modality-technical-specification-of-the-service))
- Self-Hosted Deployment: Yes — The IaaS/PaaS license installs on customer VMs (Ansible) or Kubernetes (Helm) on premises or in the customer's cloud, with an installer for setup and upgrades. ([source](https://wiki.anjanadata.com/en/solucion-tecnica/26.1/despliegue))
- Hybrid Deployment: Yes — In SaaS Hybrid mode Anjana hosts the core and the customer runs only a plugins machine next to its data systems. ([source](https://wiki.anjanadata.com/en/solucion-tecnica/26.1/saas-hybrid-mode-infrastructure-requirements-and-recommendations))
- Data Residency: Yes — SaaS runs by default in Frankfurt, Spain, Ireland, N. Virginia or N. California, and other regions are available as a paid add-on. ([source](https://wiki.anjanadata.com/en/solucion-tecnica/26.1/saas-modality-technical-specification-of-the-service))

### Open Source

- Open Source: No — Anjana Data Platform is proprietary software licensed by governed assets; no open-source edition exists. ([source](https://anjanadata.com/pricing/))

## Integrations

37 integrations, each sourced from the vendor docs.

- AI Agents: Claude, Claude Code, Cursor, GitHub Copilot
- Authentication: AWS Identities, Amazon Cognito, Auth0, Google Workspace, Keycloak, LDAP, Microsoft ADFS, Microsoft Entra, Okta, Ping
- BI: Power BI, Tableau
- Data Governance: AWS Glue Data Catalog
- Data Platform: Denodo
- Data Security: Apache Ranger
- Data Storage: Amazon S3, Azure Blob Storage, Google Cloud Storage
- Data Warehouse: Amazon Redshift, Databricks SQL, Google BigQuery, IBM DB2, Oracle DB, Snowflake, Teradata
- Databases: MariaDB, Microsoft SQL Server, MySQL, PostgreSQL
- Notifications: Google Chat, Microsoft Teams, Slack
- Query Engine: Hive SQL

## Alternatives

- [DataGalaxy](https://toolsfordata.com/tools/datagalaxy/): Consider DataGalaxy to tie data and AI products to business value.
- [Collibra](https://toolsfordata.com/tools/collibra/): Consider Collibra for a large governance program with the deepest regulatory coverage.
- [Alation](https://toolsfordata.com/tools/alation/): Consider Alation for a mature catalog with self-service discovery for business users.
- [OvalEdge](https://toolsfordata.com/tools/ovaledge/): Consider OvalEdge on a mid-market budget: governance breadth at the low end of the price range.
- [OpenMetadata](https://toolsfordata.com/tools/openmetadata/): Consider OpenMetadata for an open-source catalog with lineage, governance and data quality tests built in.
