Agentic data management
AI in every layer,
on governed data.
Validations, catalog documentation, governance and the assistant your team already uses, all working on the same governed data, inside your own environment.
- AI-assisted validationsCreate test cases from natural language, or accept AI suggestions based on data structure.
- AI catalog enrichmentGenerated descriptions, automatic field filling and suggested glossary relations.
- Governance automationsAutomated PII tagging, glossary relations and description filling, keeping critical assets validated.
- MCP context layerCatalog, lineage, glossary and quality context inside the AI tools data engineers and analysts use.
AI-assisted validations
Describe the rule, get the test. The assistant writes the SQL, suggests checks you have not thought of, and shows everything before it runs.
- Natural language to SQL: describe a data quality rule in plain words and get an executable test case
- Validation suggestions based on metadata, profiling history and the structure of the data
- Review before use: every generated test is shown in full and can be edited before it runs
- Your choice of model: OpenAI, Azure OpenAI, Anthropic, Google Gemini, or a local Ollama model, so nothing has to leave your environment
AI catalog enrichment
Documentation that keeps up with the data. Descriptions, field values and glossary links are drafted from what the platform already knows about each object.
- Generated descriptions for tables, views and columns, one at a time or as a mass action across many objects
- Context-aware suggestions that use metadata, profiling values, sample data and related glossary terms
- Suggested glossary relations and custom field values such as PII or critical data element flags
- Preview and pick: every suggestion is shown next to the current value so you apply exactly what you want
Governance automations
Write a governance rule once and let it run. New assets are tagged, linked and described as they appear, with a review step whenever you want one.
- Describe the rule once, for example “fill the PII attribute with Yes or No depending on whether the value is personal information”
- Choose the scope: which sources and objects the automation targets, and whether it applies to tables or attributes
- Review or auto-apply: accept and deny suggestions one by one, or let the automation apply them on its own
- Run on a schedule so new assets are tagged, linked and described as soon as they appear
MCP context layer
Your catalog, lineage, glossary and test results, available to the AI assistant your team already uses, with the same access rules as the rest of the platform.
- Built-in MCP server: point Claude, Cursor, VS Code or any other MCP-aware client at your SelectZero instance
- Catalog, glossary, lineage, tests and audit history become context for the assistant, through read-only tools
- Scoped to the user: the assistant sees only what the person behind the token is allowed to see
- Guided recipes for common questions, such as which tests exist for a table, where a column comes from, or what changed
See it on your own data.
Every AI feature runs inside your SelectZero deployment with your access rules in place, and with a local model no data leaves the environment.