- AI: Artificial Intelligence or AI. Anything but simplistically definable, Artificial Intelligence combines several technologies and is considered part of computer science. In principle, artificial intelligence is about artificially and intelligently reproducing human abilities in a computer or machine. This includes human abilities such as learning, pattern recognition, language comprehension, logical thinking and visual perception. AI learns skills, evolves and improves, partly with human supervision, partly unsupervised.
- APIs: Application Programming Interfaces or APIs are the connection between software and computing devices. Plug and play instead of making new connections every time.
- Classification: What document is it? Is it an invoice, is it a contract or is it an e-mail? Classification is an important step in the IDP process.
- Cloud: Cloud infrastructure now represents the most reliable medium for making continuous software improvements easily and quickly, and for scaling rapidly.
- Cognitive Document Processing: Another term sometimes used as a synonym for IDP.
- Data: The vast majority of processes in companies are based on data derived from either digital or physical documents.
- Data capture: This term is often used when speaking of Optical Character Recognition (OCR)
- Deep Learning: Deep Learning ist a subset of Machine Learning, working with neural networks. Deep Learning is the most sophisticated AI architecture.
- Digital Transformation: The transition from analog to digital processes and systems is generally referred to as digital transformation. IDP plays an important part.
- HTR: Handwritten Text Recognition (HTR) is necessary when dealing with handwritten documents and forms, which is often the case in healthcare. Advanced AI technologies help to deal with the different handwritings.
- Human-in-the-Loop: Human-in-the-loop comes into play when the AI’s confidence threshold is too low and human annotation is required. The AI actively learns with each annotation.
- Hyperautomation: While individual process automations are important in a company, they also mean only individual process optimizations. Hyperautomation, on the other hand, combines different components, tools and process automations and puts them together into one big whole, which is of high importance for companies to achieve end-to-end automation solutions.
- ICR: Intelligent Character Recognition (ICR) is used to extract handwritten text out of documents.
- Intelligent data capture: Another term sometimes used as a synonym for IDP.
- IDP: Intelligent Document Processing or IDP is used to extract unstructured data from complex documents. For this purpose, IDP systems use a variety of powerful technologies such as OCR, machine learning, deep learning, and natural language processing. Unlike traditional OCR software, IDP offers much more flexibility. IDP does not require templates and thus less configuration and is able to learn based on data.
- Intelligent Process Automation: Another term sometimes used as a synonym for IDP.
- ML: Machine Learning or ML. When talking about AI, the term Machine Learning is usually not far away either. The reason for this is that machine learning is a subset of AI. Greatly simplified, you can think of it like two circles. AI as a big picture in computer science that resembles human intelligence, at least in approach, and machine learning as an advancement that learns and improves with more and more data for specific tasks for which it was designed.
- No-Code: On a no-code platform, processes are simplified to the point where neither developers nor data scientists are needed, but the platform can be used by any employee who can operate a computer. A no-code application is more about the “what” and less about the “how.” No-code applications are becoming more common for IDP.
- NLP: Natural Language Processing or NLP is a technology used for text understanding. NLP helps the computer understand the intent and meaning of a person’s written or spoken word.
- OCR: Optical Character Recognition (OCR) is the technology used to convert data from structured documents (e.g. PDFs) into machine-readable text.
- OCR engine: The OCR engine does the actual identification of the characters and feeds back the coordinates of the found values.
- OMR: Optical Mark Recognition (OMR) is used to extract checkboxes in documents, mostly applied in surveys.
- Pre-processing: The quality of the documents is improved at the input. This includes things like noise reduction, rotation, and de-skewing.
- RPA: Robotic Process Automation or RPA is a technology for automating mundane tasks in an enterprise. When IDP is integrated, RPA functions can be greatly optimized.
- Semi-structured documents: Documents with relatively predictable information, but which vary from one company to another. An example of this would be invoices.
- Separation: When documents are received, they often arrive in batches of different documents. The separation of these documents is important here.
- Straight-Through-Processing: Straight-through processing, or STP, is possible when a lean process in an organization has no touches, no manual intervention, and runs from A to Z in a fully automated, straight-through process.
- Structured data: IDP enables the output of structured data into ERPs, DMS, CRM, etc.
- Structured documents: Official, standardized forms used for government documents such as tax forms have a structured format. This format is static and predictable within a certain framework.
- Templates: Static templates are the opposite of constantly changing document layouts and formats. With IDP, documents can be processed without templates.
- Unstructured documents: Documents that follow neither a format nor a structure and consist of text, images, and symbols. Contracts are an example of this. IDP is predestined for unstructured documents because it uses advanced AI technologies to understand the context of unstructured data, just as a human does.
- Validation: To ensure that the extracted data is correct and accurate in the first place, it must be validated. This is done by a human in the loop, where the advanced AI (machine learning) technologies learn and improve through validation.
- Variability: High document variability is common in enterprise document processes. For an intelligent automation solution like IDP, dealing with high document variability is of utmost importance.
- Workflow automation: Workflow automation is crucial for optimizing processes in companies. The cross-departmental distribution of documents and data is essential. IDP provides the high-quality data for this.