Projects
Projects are the working space for inventorying, understanding, annotating and standardizing data. Teams connect real sources, inspect table and column structures, add business meaning and organize the results into Datasets for ongoing delivery.
Establish project ownership
| Operating mode | Ownership and collaboration |
|---|---|
| Centralized | Enterprise scope, with central teams organizing governance work |
| Federated | One responsibility domain, with its teams carrying out work and central teams able to participate across domains |
Ownership defines the team's working scope and the business objects and standards available for annotation. Users enter projects through their current work identity and perform actions according to their function permissions.
Three connected workbenches
| Workbench | Main purpose |
|---|---|
| Collection | Manage connections, test sources, start collection and inspect results |
| Metadata | Browse tables, columns, constraints, indexes, lineage and structural changes |
| Annotation | Add business meaning, associate standards, handle suggestions and track completeness |
The workbenches connect around projects and sources. Users move from collection results to metadata browsing and annotation to complete a batch of governance work.
Connect sources and collect metadata
After registering connection details, test the source and start metadata collection. Collection organizes tables, views, columns and structural relationships into browsable technical metadata.
Collection runs in the background. Users can leave the page and follow progress and results through the task center. Collection history records status, duration and output for each run.
Recollection refreshes structures and records changes. Manual annotations are preserved without being overwritten by new collection results. Teams can use change information to review governance content when objects change or disappear.
Add business meaning
Annotation connects technical structures to enterprise language:
| Content | Table examples | Column examples |
|---|---|---|
| Business descriptions | Business name, definition and what one row represents | Field business name and definition |
| Business relationships | Subject, business objects, source application and terms | Data element and business terms |
| Management information | Business and technical contacts, sensitivity | Classification and masking requirements |
A supplier table can reference both Supplier and Supplier Qualification. Its registration-number column can reference the corresponding data element, while a contact-number column carries sensitivity and masking requirements.
Enterprise projects use enterprise standards. Domain projects in Federated mode can use enterprise or same-domain standards and business objects. Annotation choices reflect the project scope to keep references consistent.
Track annotation progress
The annotation workbench shows table and column completeness, helping teams find unannotated, partially completed and review-needed content. Users can inspect missing items and work through tables in sequence.
Completeness considers the table's business name, definition, grain, subject and sensitivity; and each column's business name, definition, effective classification and masking rule when sensitive.
Use AI-assisted annotation
Users describe a governance task to an AI agent, which uses table and column structures, existing standards, enterprise knowledge and data dictionaries to recommend names, definitions, classifications and related annotations. Specialist agents can collaborate on complex tasks.
Results appear as suggestions for users to inspect, adjust, accept or reject. Suggestions become official annotations only after human review and acceptance; unaccepted suggestions do not count as completed governance work. Direct manual maintenance and AI assistance work together in the same process.
Apply standards and build datasets
After associating a column with a data element, teams can compare its definition, type, length and value range with the standard to identify data or definition issues.
For ongoing sharing, select a group of project tables to create a dataset, describe its output ports and contracts, and associate implementations in the appropriate environments. The project organizes the work; the dataset maintains consumer commitments and versions over time.
View project activity and history
Project activity brings together source changes, collection runs, annotation edits and suggestion handling. Project history records changes to names, descriptions and ownership, making adjustments to collaboration boundaries traceable.
Activity and history respect the user's project scope and function permissions. Records link to the relevant workbench so teams can inspect the objects and governance results.
Standards
Align data definitions, controlled values, business measures and security requirements so teams and systems share understandable, reusable standards.
Datasets
Organize governance outputs into datasets with clear ownership, ports, contracts, versions and environment implementations so consumers know what data they can use.