Documents become usable information.
One of the key tasks in document digitization is not only to capture the content digitally but also to understand it automatically and make relevant information available for downstream business processes. To this end, documents must be recognized and classified, and the required information reliably extracted.
The requirements vary depending on the document type:
Structured documents
Forms, for example, have a largely fixed structure and fixed positions for specific information.
Semi-structured documents
Invoices, delivery notes, or purchase orders contain comparable information, though its position and presentation may vary from document to document.
Unstructured documents
Cover letters, contracts, or other correspondence where relevant information can only be identified from the context of the content.
Many years of experience in intelligent document recognition
For many years, we have been working on the automated recognition, classification, and data extraction of documents. We have accompanied different generations of capture and recognition technologies – from earlier solutions and technologies such as DoKuStar and OpenText Capture Center (OCC) to today’s intelligent capture platforms.
With OpenText Capture—formerly OpenText Intelligent Capture—modern methods are now available that combine traditional OCR technologies with machine learning and artificial intelligence. This enables documents to be automatically classified and relevant information to be extracted in a targeted manner.
Recognize – Classify – Extract – Validate
A typical intelligent capture process comprises several steps:
1. Recognition
Incoming paper or electronic documents are analyzed and prepared for further processing.
2. Classification
The system automatically recognizes the type of document—e.g., invoice, delivery note, contract, or form.
3. Extraction
The required information is extracted from the document—such as the invoice number, order number, or date in the case of an invoice.
4. Validation
Recognized information can be verified and corrected using rules, existing master data, or—if necessary—by a user.
5. Handover
Subsequently, the document and data are automatically transferred to the downstream ERP, document management, archiving, or workflow system.
Artificial intelligence expands the possibilities.
Modern AI technologies are fundamentally transforming traditional document recognition.
While traditional capture solutions often relied on rigidly defined rules, positions, and trained document classes, machine learning and new AI techniques increasingly enable the processing of complex and variable documents as well.
This makes Intelligent Capture a crucial link between incoming documents and automated business processes. Our goal is not to deploy as much AI as possible, but rather to select the right combination of proven document recognition, rules, machine learning, and modern AI techniques for each use case.
Experience determines the level of automation.
High-performance capture technology alone does not make for a successful process.
The decisive factor is how document classification, data extraction, validation, and integration into downstream systems are aligned with one another.
Drawing on our extensive experience with document recognition and document-based business processes, we help our customers extract structured, reliable information from incoming documents for automated downstream processing. This transforms a scanned or electronically received document into a usable component of a digital business process.
More information
OpenText Intelligent Classification
Use natural language processing capabilities to analyze unstructured content and process the information it contains.
OpenText Continuous machine learning: Your AI edge
How OpenText leverages a continuous machine learning (CML) approach that offers flexibility, accuracy, and efficiency while minimizing or eliminating manual model retraining.