Prodigy - OpenText File 360 Integration and Automation

Common Integration Use Cases Between Prodigy and OpenText File 360

Prodigy and OpenText File 360 complement each other well in enterprise AI programs where sensitive source data must be shared, reviewed, and governed before or during annotation. OpenText File 360 provides secure, auditable file access and collaboration, while Prodigy enables efficient labeling and active learning for machine learning teams. Together, they support controlled data movement, cross-functional review, and compliant model development workflows.

• Secure delivery of raw training data to annotation teams

  Flow: OpenText File 360 to Prodigy

  Business units can store source documents, images, or text datasets in OpenText File 360 and grant Prodigy users controlled access to the exact files approved for labeling. This avoids emailing datasets or using consumer file-sharing tools, while preserving audit trails and access restrictions for regulated data.

  Value: Faster dataset onboarding, reduced security risk, and better governance over who can access training data.

• Controlled sharing of labeled datasets for model training and review

  Flow: Prodigy to OpenText File 360

  After annotation is completed in Prodigy, labeled exports can be automatically written to OpenText File 360 for review by data science leads, compliance teams, or external partners. This creates a governed handoff point for approved training sets before they are used in TensorFlow, PyTorch, or downstream MLOps pipelines.

  Value: Centralized storage of approved datasets, easier review cycles, and stronger traceability for model inputs.

• Human-in-the-loop review of sensitive annotations

  Flow: Prodigy to OpenText File 360 and back to Prodigy

  For high-risk use cases such as legal documents, customer communications, or medical records, annotators can export disputed or low-confidence items from Prodigy into OpenText File 360 for secure review by subject matter experts. Reviewers can return approved files or corrected labels to Prodigy for finalization.

  Value: Better label quality, structured expert review, and reduced rework on sensitive datasets.

• Governed collaboration with external labeling vendors

  Flow: OpenText File 360 to Prodigy

  Organizations can use OpenText File 360 to share only the required subset of files with external annotation vendors while maintaining access controls, expiration policies, and audit logs. Vendors work in Prodigy on the assigned data, and only approved outputs are returned to the enterprise environment.

  Value: Safer outsourcing of labeling work, tighter vendor control, and improved compliance with data handling policies.

• Audit-ready dataset lineage for regulated AI projects

  Flow: Bi-directional

  OpenText File 360 can serve as the system of record for dataset versions, approvals, and file access history, while Prodigy records annotation activity and label changes. Together, they provide a more complete lineage trail showing where the data came from, who accessed it, and how it was labeled before model training.

  Value: Stronger auditability for regulated industries such as financial services, healthcare, and public sector AI initiatives.

• Exception handling for low-confidence or ambiguous samples

  Flow: Prodigy to OpenText File 360

  When annotators encounter ambiguous samples in Prodigy, those files can be routed to OpenText File 360 for escalation to legal, compliance, or domain experts. Once a decision is made, the approved interpretation can be returned to Prodigy to update the label set and maintain consistency across the project.

  Value: Faster resolution of edge cases, fewer labeling inconsistencies, and better model performance on complex scenarios.

• Secure distribution of model training packages to downstream teams

  Flow: Prodigy to OpenText File 360

  Completed annotation packages, including labeled files, instructions, and review notes, can be published to OpenText File 360 for downstream data science, QA, or MLOps teams. This gives stakeholders a controlled way to retrieve the exact dataset version used for training or validation.

  Value: Simplified handoff between teams, reduced version confusion, and improved reproducibility of model training runs.