Kenseme uses AI in many places, from drafting a description to generating a whole ontology. AI only runs when you start it. Most features propose and wait for you to accept; a few, such as bulk description generation or asking the Ontology Assistant to make an edit, write the result straight away because that’s what you asked for. This page lists every AI feature by where it lives in the app.
Every AI feature draws on your organization’s monthly AI usage allowance. See AI usage allowance.
Business context
| Feature |
What it does |
Learn more |
| Glossary Extractor |
Reads a document or pasted text and suggests glossary terms and definitions for you to pick from. Open it with Glossary Extractor on a glossary. |
Extract glossary terms |
| Process document processing |
Summarizes uploaded process documents and indexes them so you can search them by meaning and use them as sources for ontology generation. A document’s Guidelines tell the AI how to read it. |
Upload and search process documents, Work with a process document |
| Industry standard grounding |
Uses an industry standard’s vocabulary as a source when the AI generates an ontology. |
Search and use industry standards |
Data context: ontologies
| Feature |
What it does |
Learn more |
| Ontology Generator |
An agent builds a new ontology, or extends an existing one, from sources you pick: schemas, databases, a lakehouse, process documents, glossaries, and standards. You review its proposal item by item before anything is saved. Draft with AI helps you write the instructions. |
Generate an ontology with AI |
| Ontology Assistant |
A chat drawer for asking about an ontology or having it add, rename, and remove items for you, with undo. It has two modes: General, and Tool Assisted, which is available once the ontology has been generated. |
Chat with the Ontology Assistant |
| Ontology descriptions |
Add Descriptions writes descriptions for every class, relationship, and attribute at once. In the editor, Auto Generate Description writes one for a single item, and a class can have descriptions written for all its relationships or attributes. |
Generate descriptions for an ontology |
| Import warning explanations |
Explain these warnings explains, in plain language, the warnings from an ontology file import. |
Create or import an ontology |
| Lakehouse binding suggestions |
In the binding workspace’s More menu, Generate suggestions proposes which lakehouse tables and columns map to each class, attribute, and relationship; Generate relationship suggestions proposes joins for relationships only. |
Bind an ontology to a lakehouse |
| Database binding suggestions |
The same two suggestion actions for ontologies bound to a relational database. |
Bind an ontology to a database |
| Version notes |
Draft with AI writes a description of what changed when you create an ontology version. |
Ontology versioning |
| Deploy and sync risk summary |
AI risk summary explains what a version deploy or Fabric sync is about to change and what could go wrong. |
Ontology versioning, Sync an ontology with Fabric |
Data model: schemas
| Feature |
What it does |
Learn more |
| Schema Agent |
A chat on the schema’s Schema Agent tab. Ask about the design, and it can propose new views, procedures, functions, indexes, and tables, or changes to existing ones. Nothing changes until you click Create or Apply on its proposal. |
Work with the Schema Agent |
| Query Agent |
Turns a plain-English question into SQL for you to copy and run yourself, from the schema’s definition alone. |
Generate a query with AI |
| Schema import |
Splits and parses DDL files, explains import failures, and writes descriptions for imported objects. |
AI in schema import |
| Descriptions |
Generates descriptions for tables, views, procedures, and functions, one at a time or across a schema. |
AI in schema import |
| Name suggestions |
Suggests table and column names that follow your naming conventions. |
AI in schema editing |
| SQL object analysis |
Works out what views, procedures, functions, and scripts depend on and do. |
AI in SQL object analysis |
| Script validation |
Validate script with AI agent reviews a generated create or migration script for problems specific to the target dialect. |
Generate a create script |
| Dialect conversion |
When you duplicate a schema into another dialect, converts procedure bodies, views, and defaults for the target. |
Create a schema |
| Star Schema designer |
Run AI Analysis proposes dimensions and facts, then drafts source views, tables, and load procedures. |
AI in the Star Schema designer |
Data model: Pipelines
| Feature |
What it does |
Learn more |
| AI rename |
Run AI Rename suggests readable destination column names for cryptic source names. |
AI in Pipelines |
| Expression check |
Checks view-based expressions against the target dialect. |
AI in Pipelines |
Agents
| Feature |
What it does |
Learn more |
| Data Agent |
Ask for data in plain English against an ontology bound to a database. It writes and runs a read-only query, shows the rows, the SQL, and its assumptions, and can chart the result. It also suggests questions to start from. Needs the Data Agent feature. |
Ask the Data Agent |
| Ontology Data MCP |
Lets an outside AI client ask business questions of an ontology, answered from its database bindings. Needs the Ontology Data MCP feature. |
Expose an ontology over MCP |
What AI in Kenseme doesn’t do
- It doesn’t act on its own. Nothing runs until you start it. Suggestions wait for you to accept, create, or apply them. When you ask the Ontology Assistant to edit, it asks before destructive changes, and every edit can be undone.
- It doesn’t write to your databases. The Data Agent and Ontology Data MCP run read-only queries. Everything else works on the definitions stored in Kenseme.
- It doesn’t run your generated SQL. You deploy scripts and procedures yourself.
