Sample deliverables
Inspect synthetic public examples of the guidelines, schemas, QA evidence, redaction records, evaluation reports, and acceptance summaries used in controlled Hebrew AI workflows.
What it demonstrates: Shows label definitions, edge cases, reviewer instructions, and acceptance rules.
Intended use case: Hebrew text classification and annotation calibration.
Quality rules: Explicit fields, acceptance criteria, reviewer notes, and documented limitations.
Limitations: Public demonstration only; not a client deliverable and not evidence of performance for a specific production system.
Download PDFWhat it demonstrates: Shows structured fields, allowed values, validation rules, and examples.
Intended use case: Dataset collection, exchange, and validation.
Quality rules: Explicit fields, acceptance criteria, reviewer notes, and documented limitations.
Limitations: Public demonstration only; not a client deliverable and not evidence of performance for a specific production system.
Download CSVUse the public samples to evaluate the documentation approach, then request a controlled pilot based on your own data, model, and acceptance criteria.
Request a PilotWhat it demonstrates: Shows accepted and rejected counts, defect types, severity, and corrective actions.
Intended use case: Batch acceptance and vendor-delivery review.
Quality rules: Explicit fields, acceptance criteria, reviewer notes, and documented limitations.
Limitations: Public demonstration only; not a client deliverable and not evidence of performance for a specific production system.
Download PDFWhat it demonstrates: Shows field-level redaction decisions without exposing real personal data.
Intended use case: Privacy-aware document and email processing.
Quality rules: Explicit fields, acceptance criteria, reviewer notes, and documented limitations.
Limitations: Public demonstration only; not a client deliverable and not evidence of performance for a specific production system.
Download CSVUse the public samples to evaluate the documentation approach, then request a controlled pilot based on your own data, model, and acceptance criteria.
Request a PilotWhat it demonstrates: Shows Hebrew language, factuality, instruction, safety, and consistency scoring.
Intended use case: Hebrew LLM and chatbot evaluation.
Quality rules: Explicit fields, acceptance criteria, reviewer notes, and documented limitations.
Limitations: Public demonstration only; not a client deliverable and not evidence of performance for a specific production system.
Download PDFWhat it demonstrates: Shows structured error categories, severities, and remediation fields.
Intended use case: LLM evaluation, red-teaming, and QA.
Quality rules: Explicit fields, acceptance criteria, reviewer notes, and documented limitations.
Limitations: Public demonstration only; not a client deliverable and not evidence of performance for a specific production system.
Download JSONUse the public samples to evaluate the documentation approach, then request a controlled pilot based on your own data, model, and acceptance criteria.
Request a PilotWhat it demonstrates: Shows reading-order, character, table, field, and layout defects.
Intended use case: Hebrew OCR and document AI evaluation.
Quality rules: Explicit fields, acceptance criteria, reviewer notes, and documented limitations.
Limitations: Public demonstration only; not a client deliverable and not evidence of performance for a specific production system.
Download PDFWhat it demonstrates: Shows batch disposition, rejection reasons, thresholds, and release decision.
Intended use case: Data collection and recurring production batches.
Quality rules: Explicit fields, acceptance criteria, reviewer notes, and documented limitations.
Limitations: Public demonstration only; not a client deliverable and not evidence of performance for a specific production system.
Download PDFUse the public samples to evaluate the documentation approach, then request a controlled pilot based on your own data, model, and acceptance criteria.
Request a Pilot