Skip to main content
ABBYY Vantage is an intelligent document processing (IDP) platform that lets organizations extract structured data from a wide variety of documents (structured, semi‑structured, unstructured). A core component is the Skill Designer, which lets users create, train, and deploy what ABBYY calls Skills — models/rules for classifying documents, extracting data, and integrating into business/process workflows. In short, the Skill Designer allows you to build “document skills” (document type definitions + extraction rules), “classification skills” (to detect which type of document you’ve got), and “process skills” (to put together multi‑step logic combining classification, document skills, business rules).

​
Types of Skills in the Skill Designer

You can work with the following Skills in the Skill Designer:
  • Classification skills
  • Document skills
  • OCR skills
  • Process skills

​
How it Works

Here’s a high‑level flow for using Skill Designer:
  1. Define/document sample set. Collect sample documents of the types you want to process (structured, semi‑structured, unstructured).
  2. Classification (if needed). If you have multiple document types, build a classification skill to auto‑route documents to correct handling.
  3. Label / annotate fields. Define which pieces of data you want extracted. This involves marking fields on sample documents.
  4. Train the document skill. Use ABBYY’s model/training pipeline (fast learning or deep learning depending on complexity) to train the extraction model.
  5. Review & validate. Use human review / validation to catch mis‑extractions or confidence issues. Adjust labeling, business rules, etc., to improve accuracy.
  6. Publish / deploy. Once satisfied with performance, publish the skill so it can be used within business workflows. Skills are made discoverable and can be called via APIs.
  7. Monitor & iterate. Track performance metrics. Over time feed new samples to retrain or adjust the skill to handle more document variation.