Prodigy - OpenText Content Metadata Service - Dictionary Integration and Automation

Common Integration Use Cases Between Prodigy and OpenText Content Metadata Service - Dictionary

1. Standardized annotation schema for AI training datasets

Data flow: OpenText Content Metadata Service - Dictionary ? Prodigy

Use the OpenText dictionary as the master source for labels, categories, and controlled vocabularies used in Prodigy projects. This ensures that annotation teams apply consistent metadata across image, text, and document datasets, especially when multiple business units contribute to model training.

Business value: Reduces labeling inconsistency, improves model quality, and avoids rework caused by mismatched taxonomies.

2. Governance of training data classification across departments

Data flow: Bi-directional

When Prodigy annotators create or refine labels during active learning, those new or updated terms can be reviewed and synchronized back to the OpenText dictionary for governance approval. Once approved, the updated schema is redistributed to Prodigy and other content platforms.

Business value: Keeps AI labeling aligned with enterprise information governance standards while allowing data science teams to evolve schemas quickly.

3. Controlled vocabulary enforcement for document and text annotation

Data flow: OpenText Content Metadata Service - Dictionary ? Prodigy

For NLP use cases such as contract review, customer correspondence classification, or case file tagging, Prodigy can pull approved metadata values from OpenText to constrain annotators to valid terms only. This is especially useful for regulated industries where classification must match enterprise records policies.

Business value: Improves compliance, searchability, and downstream reporting consistency.

4. Metadata-driven sample selection for active learning

Data flow: OpenText Content Metadata Service - Dictionary ? Prodigy

OpenText metadata definitions can be used to filter and prioritize source content before it enters Prodigy. For example, only documents tagged as high-risk, customer-facing, or region-specific can be selected for annotation to train specialized models.

Business value: Focuses labeling effort on the most valuable content and accelerates model performance for targeted business scenarios.

5. Annotation output mapped back to enterprise content repositories

Data flow: Prodigy ? OpenText Content Metadata Service - Dictionary

After annotation, Prodigy outputs can be transformed into standardized metadata values and written back to OpenText-managed repositories. This enables enriched documents, images, or records to carry AI-derived classifications that can be used for search, retention, routing, or audit workflows.

Business value: Turns model training outputs into operational metadata that improves enterprise content management and automation.

6. Cross-system metadata validation before publishing AI labels

Data flow: Bi-directional

Before labels from Prodigy are accepted into production workflows, they can be validated against the OpenText dictionary to confirm data type, allowed values, and hierarchy rules. Any exceptions can be flagged for review by information governance or domain experts.

Business value: Prevents invalid metadata from entering downstream systems and reduces governance risk.

7. Shared taxonomy for multimodal AI initiatives

Data flow: OpenText Content Metadata Service - Dictionary ? Prodigy

Organizations building multimodal AI models can use the OpenText dictionary to define a common taxonomy across images, text, and scanned documents. Prodigy then applies that taxonomy consistently during annotation, enabling unified training data across content types.

Business value: Supports enterprise-wide AI programs with a single metadata model, improving reuse and interoperability across teams.

8. Metadata change management for evolving AI programs

Data flow: Bi-directional

As business rules change, OpenText can publish updated dictionary versions that Prodigy consumes for new annotation tasks. In parallel, Prodigy can report label usage patterns, missing categories, or ambiguous terms back to OpenText so metadata stewards can refine the enterprise schema.

Business value: Creates a closed-loop process between AI teams and content governance teams, helping organizations keep metadata models current and operationally useful.