In the app: Data Context

Data Context

Data Context is home to ontologies — a formal description of the things your business deals with (customers, orders, stores) and how they relate. Build one by hand, import one, or generate one with AI; then bind it to lakehouse or database tables so the Data Agent and other AI tools can answer questions against your real data.

Ontologies

Ontologies overview What an ontology is in Kenseme, what it is made of, and where to go to create, edit, bind, version, and share one. Create or import an ontology Start a new ontology with the Add Ontology wizard: blank, from a Turtle or OWL file, from Microsoft Fabric, generated by AI, or from an industry standard. Generate an ontology with AI Point an AI agent at your schemas, databases, lakehouse, documents, glossaries, or standards, then review its proposed ontology before anything is saved. Duplicate an ontology Copy an ontology’s classes, relationships, attributes, and prefixes into a new ontology, and optionally copy its data bindings too. Edit an ontology Add, rename, and delete classes, relationships, attributes, and instances in the ontology workbench. Changes save as you make them. Visualize the ontology diagram See an ontology as a graph of classes and relationships, focus on one neighborhood, and switch to Edit mode to change the model on the canvas. Generate descriptions for an ontology Have AI write descriptions for one ontology item, for the relationships or attributes of a class, or for the whole ontology at once. Chat with the ontology assistant Ask the AI assistant about an ontology or have it add, rename, and remove items for you, then review and undo its changes on the Activity tab. Link to an ontology entity Every class, relationship, attribute, and instance has its own address. Copy it to share, bookmark, or open an entity in a new tab. Ontology versioning Snapshot an ontology with its bindings as an immutable version, inspect past versions read-only, export them, and deploy one to Microsoft Fabric. Export an ontology Download an ontology as a Turtle or OWL (RDF/XML) file to back it up, share it, or open it in another ontology tool.

Connect to data

Bind an ontology to a lakehouse Map an ontology’s classes, attributes, and relationships onto tables and columns in a Microsoft Fabric lakehouse, with AI suggestions and validation. Bind an ontology to a database Map an ontology’s classes, attributes, and relationships onto an external database’s tables and columns so the Data Agent can query live data. Sync an ontology with Microsoft Fabric Publish an ontology and its lakehouse bindings to a Microsoft Fabric ontology item, pull Fabric-side changes back, preview every run, and resolve conflicts. Ontology sync history See where an ontology came from and every push to and pull from Microsoft Fabric since, with what each run changed, who ran it, and why a run failed.

Share over MCP

Expose an ontology over MCP Turn on an ontology’s MCP server so an AI assistant can ask business questions about your data, choose whether it gets rows or only SQL, and connect a client. Control what the ontology MCP can see Choose what an ontology’s MCP server reveals: discovery tools, redacted classes and attributes, stricter per-person limits, and an audit of every call. Data MCP tools The fifteen tools an ontology’s MCP server offers an AI assistant: what each returns, which use your AI allowance, and which setting turns each on.