Reference
Kenseme glossary
A quick reference to the terms you’ll meet in Kenseme, in alphabetical order. Each links to the page that explains it in full.
| Term | Definition |
|---|---|
| Activity feed | The list of recent changes across your organization on Home and the Activity page. See Activity feed. |
| AI usage allowance | The amount of AI work your plan includes each month, measured in usage units. Usage packs top it up. See AI usage allowance. |
| Assumptions | The choices the Data Agent made that you didn’t state, such as a time range, shown next to each answer so you can change them. See Ask the Data Agent. |
| Attribute (governance) | A custom metadata field with a type, such as Owner or Sensitivity, that you set on schema objects, ontology entities, and glossary terms. See Custom attributes overview. |
| Attribute (ontology) | A property of an ontology class that holds a value, such as a customer’s name. Called a Data Property in OWL and a Field in Microsoft Fabric. See Ontologies overview. |
| Attribute group | A named bundle of custom attribute definitions. See Create attribute groups. |
| Background job | Long-running work, such as an import or AI generation, that runs on the server while you carry on. See What runs in the background. |
| Base IRI | The web-style address prefix that an ontology’s identifiers start with. See Ontology settings. |
| Binding | The link between an ontology entity and the physical data behind it: a class to a table, an attribute to a column, a relationship to a join. See Bind an ontology to a lakehouse and Bind an ontology to a database. |
| Binding environment | One named target (Dev, Test, Prod, and so on) that an ontology’s bindings point at. |
| Bus matrix | The grid in the Star Schema designer showing which dimensions apply to which facts. |
| Class | A kind of thing in an ontology, such as Customer or Order. Called an Entity in Microsoft Fabric. |
| Computed expression | A SQL expression in a Pipeline that derives a destination column from one or more source columns. |
| Data Agent | The agent that answers plain-English questions from your data by writing and running read-only queries over a database-bound ontology. See Data Agent overview. |
| Database connection | Saved details for reaching an external database, used for imports, ontology generation, and database binding. See Database connections overview. |
| Dialect | One of the eight SQL variants Kenseme generates for. See Supported database targets. |
| Dimension / fact | The two building blocks of a star schema: dimensions describe (Customer, Date), facts measure (Sales). See Star schema overview. |
| Draft (Ontology Generator) | The hidden copy the Ontology Generator writes its proposal into. You review it, and only the parts you accept reach your ontology. See Generate an ontology with AI. |
| Drift | Differences between a star schema plan and the current columns of its source tables. See Handle star schema source drift. |
| Environment (schema) | A schema’s mapping to a real database and schema name, used to generate environment-specific SQL. See Schema settings. |
| Feature flag | A switch that turns an optional feature on for an organization. See Feature flags. |
| Finalize (Pipeline) | Commits a Pipeline, creating its tables, columns, and procedures in the destination schema. See Finalize a Pipeline. |
| Glossary | A container of business terms and their definitions. See Business glossaries overview. |
| Industry standard | A published vocabulary such as GS1 or UNSPSC that you can search and use as a source for ontologies. See Industry standards overview. |
| Instance | A specific named member of a class, such as Canada as an instance of Country. Called a Named Individual in OWL. |
| Label | A colored chip you attach to schemas for quick visual grouping. See Create and apply organization labels. |
| Lakehouse | A Microsoft Fabric store that holds data as Delta tables. Ontologies can bind to one, and Kenseme can generate PySpark notebooks for it. See Bind an ontology to a lakehouse. |
| Load strategy | How a Pipeline moves rows into a target table: full reload, incremental, slowly changing dimension, or fact. See Pipeline load strategies. |
| Mapping (Pipeline) | The pairing of a source column with a destination column in a Pipeline. See Map source to target columns. |
| MCP | Model Context Protocol, the standard that lets outside AI clients use a schema or ontology you expose from Kenseme. See Expose a schema over MCP and Expose an ontology over MCP. |
| Migration script | SQL that takes a database from one schema version to another. See Generate a migration script. |
| Ontology | A model of your business: the kinds of things that exist (classes), how they relate (relationships), what describes them (attributes), and named examples (instances). See Ontologies overview. |
| Ontology Assistant | The chat drawer that answers questions about an ontology and edits it for you, in General or Tool Assisted mode. See Chat with the Ontology Assistant. |
| Ontology Generator | The AI agent that builds or extends an ontology from sources you choose. See Generate an ontology with AI. |
| Ontology version | A saved, read-only snapshot of an ontology and its bindings, which you can deploy to Microsoft Fabric. See Ontology versioning. |
| Organization | The top-level container for everything you build in Kenseme, and the people who build it. See What an organization is. |
| Pipeline | A configured flow of data from source tables to target tables across schemas, with generated load procedures. See Pipelines overview. |
| Process document | An uploaded document describing how your business works, indexed so you and the AI can search it. See Process documents overview. |
| Procedure | The generated code that moves data: a stored procedure, or a PySpark notebook on Databricks and Fabric Lakehouse. |
| PySpark notebook | The output format for Databricks and Fabric Lakehouse targets. See PySpark notebooks. |
| Query Agent | The schema tab that writes SQL from a plain-English question for you to run yourself. See Generate a query with AI. |
| Read-only organization | An organization in which nobody can make changes, because of its billing status or an administrator’s action. Not the same as the Read Only role. See Account states and read-only mode. |
| Relationship | A named link between two ontology classes, such as placesOrder. Its identifier is always lowerCamelCase. Called an Object Property in OWL. See Edit an ontology. |
| Role | What a member can do in an organization: Admin, Read/Write, Read Only, or None. See Roles in Kenseme. |
| Schema | A logical container of tables, views, procedures, functions, indexes, relationships, and settings. See Schemas overview. |
| Schema Agent | The chat on a schema’s Schema Agent tab that answers design questions and proposes new objects and changes. See Work with the Schema Agent. |
| Schema version | A frozen snapshot of a schema at a point in time. See How versioning works. |
| Seat | One member place in an organization. Active members and pending invitations each use a seat. See Manage users in your organization. |
| Site Admin | Macula staff who operate Kenseme and use System Administration. See System administration. |
| Star schema plan | A Kimball-style dimensional model designed from a source schema. See Star schema overview. |
| Sync (Fabric) | Pulling an ontology and its bindings from, or pushing them to, a Microsoft Fabric ontology item. See Sync an ontology with Fabric. |
| Term | A single entry in a glossary. See Define a term. |
| Usage pack | A one-off top-up to the AI usage allowance that rolls over until used. See AI usage allowance. |