Generative AI is a hot topic these days, but it can be tough to figure out how it adds value to your business. Even after starting to research it, you may end up with more questions than answers.
To bridge this gap, DocuWare sponsored the Market Momentum Index: Intelligent Document Processing (IDP) Survey 2025, conducted by AIIM and Deep Analysis. This survey dives into intelligent document processing (IDP), which leverages AI to understand text in context, classify it, and extract data from unstructured and structured documents. Implementing ID quickly shows value from an investment in AI while other AI initiatives take more time. IDP can be a standalone product or part of a larger software solution; this survey applies to both.
AIIM collected data from 600 companies in the US, Germany, Austria, and Switzerland. These companies, from industries like finance, manufacturing, healthcare, and insurance, all have revenues of over $10 million and more than 500 employees.
Table of Contents
- How intelligent document processing (IDP) provides an AI-driven competitive edge
- What is IDP?
- IDP adoption and use cases
- Success stories: Transforming workflow automation with DocuWare
- Why intelligent document processing is essential for SMBs
- Future outlook and the way forward
How intelligent document processing (IDP) provides an AI-driven competitive edge
This in-depth survey shows a big jump in AI adoption and expanding IDP uses, while also highlighting the importance of external users and data security challenges.
66% of new IDP projects will replace an existing system. Innovative startups and scale ups will continue to take market share, while legacy vendors loaded down by technical debt will struggle to catch up. Proof-of-concept testing and industry-specific expertise remain critical for buyers, emphasizing the need for vendors to demonstrate tangible value.
78% of companies are using AI now — a massive jump from just a year ago, when stalled projects and concerns about the risk profile of AI caused many to predict this was just an exaggeration that would amount to a small amount of actual change. The data now suggests that the trend has increased dramatically – at least for IDP projects.
A major shift from IDP’s traditional invoice processing use case: Licenses and permits, know your customer (KYC), contracts/agreements, and human resource files are all on the rise. This is proof that IDP has expanded rapidly beyond its traditional back-office functions and into front-office, customer-facing operations.
External Users Surge: 62% of IDP systems now involve external users, confirming the significant shift toward front-office applications;
Data Security & Privacy is the #1 Challenge: The biggest hurdle for IDP implementation is the security and privacy of the data that will be processed.
Massive Growth and Disruption Driven by GenAI
Many executive teams have been doubtful about the cost vs. benefit equation when it comes to AI. Until recently some thought it was just another overhyped trend. Now, the data shows a noticeable change, particularly for intelligent document processing initiatives.
65% of respondents said they have or are considering a new project, showing a massive increase in IDP technology. Of that group, 45% plan to automate a new use case, an extraordinary result. This finding supports the premise that GenAI capabilities are unleashing new use case potential, especially in front-office or customer-facing applications.
Paper’s Persistent Role in IDP Processes
Paper is still incorporated into IDP processes. 61% of IDP processes still include paper documents and 48% of respondents expect paper volumes to increase next year.
Why is paper still such a big part of AI-driven document processing, and why is it so tough to eliminate? The reasons vary from concern about the cost of digital transformation to reluctance to change existing processes. The key takeaway is this: digitizing and managing paperwork is still a vital part of using most IDP systems and will continue to be important for the foreseeable future.
Surprisingly, fax machines are still used in 37% of IDP processes today. This statistic is consistent across all regions, debunking the common belief that the US healthcare system is the primary large-scale user of faxed documents. The top users of fax are government agencies and health insurance companies.
What is IDP?
Industry experts estimate that most business data—between 80% and 90%—is unstructured. Not using this data is a lost chance for valuable insights. Intelligent document processing uses AI to address this challenge.
IDP software uses AI to enable computers to read and extract data from paper and electronic documents, serving as an update to older software like traditional OCR. The software captures structured and unstructured data accurately using machine learning (ML), natural language processing (NLP), deep optical character recognition (OCR), and other AI tools. IDP interprets document content, not just scanning, sorting, and extracting data from it.
Structured data, like Excel spreadsheets, time sheets, customer contact info and product details, has a fixed format. Unstructured data lacks a structured format, such as emails, social media posts, images, audio files, handwritten notes and contracts.
IDP can be either a standalone product or part of a broader software suite, and this survey covers both options.
IDP Market Insights
Find out how companies like yours are using AI to drive efficiency and gain a competitive advantage.
DocuWare IDP’s advanced features
- Pre-processing
- Document Splitting: The software automatically separates large files or batches into individual files without requiring separator sheets or barcodes.
- Precision Cropping: Ensures scanned documents are properly aligned and cropped neatly.
- Automated Data Retrieval: IDP systems use AI and machine learning algorithms to pull out data. AI can grasp language contextually, allowing accurate extraction from both structured and unstructured documents.
- Indexing: IDP recognizes different document formats and categorizes them based on their content and characteristics. It can also trigger specific workflows depending on the document type.
- Deep optical character recognition (OCR): This advanced form of OCR uses deep learning, a type of AI, to improve text recognition and data extraction. OCR converts scanned documents or images into digital text. It helps scanners turn typed content into text that computers can process. This digital text can be indexed, validated, routed, or used in other systems, eliminating the need for manual retyping. OCR works best with structured and clearly printed text.
- Handwritten text recognition (HTR): While OCR works well for printed or typed text with standard characters, HTR handles the variations and nuances found in human handwriting. The primary purpose of HTR is to enable handwritten text, such as annotations, letters, insurance forms and onboarding documents to be completely searchable and easy to access.
- Flexible AI models: Machine learning enables IDP software to learn from past documents without needing extra coding. This involves pre-set or user-tailored AI models. While pre-configured models are easy to implement and save time and effort, custom algorithms cater to your company's unique needs, trained specifically with your data and aligned with your goals.
- Data validation: IDP ensures the accuracy of extracted data using rules-based algorithms right during the extraction process, avoiding any need for a separate check.
Technologies powering IDP: A definition of terms
- ArtificiaI intelligence: This is a branch of computer science aimed at creating intelligent machines that can perform tasks requiring human-like thinking, such as learning, reasoning, and solving problems.
- Machine learning: A subset of AI, machine learning uses algorithms trained on data to build models that can perform tasks like image categorization, data analysis or predicting price changes.
- Deep optical character recognition (OCR): Unlike traditional OCR, which follows predefined rules, Deep OCR models are trained on large datasets to understand and process text in unstructured documents like meeting minutes and emails. The result? OCR technology that can handle more complex documents with more accuracy than traditional methods.
- Neural networks: Part of machine learning, neural networks mimic the structure and function of the human brain. They consist of interconnected nodes structured in layers that transmit and process data. These networks form the basis of deep learning and can learn patterns from data to make predictions or decisions.
- Natural language processing (NLP): NLP specifies how computers understand, process, and manipulate human languages. It includes interpreting language meanings, translating languages, recognizing language patterns using statistical methods, machine learning, neural networks, and text mining. Tools like Google Translate and automated chatbots use NLP.
- Handwriting recognition (HTR): HTR finds handwritten text in documents and converts it into editable text. It can be used to digitize annotations on printed text, patient notes, tax documents, loan applications, and in many other cases that involve handwritten information.
- Traditional OCR: This type of OCR converts scanned documents or images into digital text. It helps scanners turn typed content into text that computers can process. This digital text can be indexed, validated, routed, or used in other systems, eliminating the need for manual retyping. OCR works best with structured and clearly printed text.
IDP adoption and use cases
Current and Future IDP Projects
Survey findings revealed two unexpected trends: Human resource files and contracts/agreements have become as popular for IDP as invoices and financial statements. These long, unstructured documents used to be tough to process with machine learning or templates alone.
Over the next two years, the processing of other document types such as licenses and permits, know your customer (KYC) and customer, claims forms, and customer and employee onboarding documents are expected to grow much faster than invoices and financial statements. According to the AIIM survey, IDP will continue to expand beyond its usual back-office roles into more customer-facing, front-office functions.
Two-year growth projections for IDP projects
| Licenses and permits | 54% |
| Know your customer | 29% |
| Claims forms | 27% |
| Loan origination forms | 16% |
| Correspondence | 14% |
| New customer onboarding | 12% |
| Medical records | 10% |
| Contracts | 10% |
| Receipts | 8% |
| Invoices and/or purchase orders | 6% |
| Identification documents | 4% |
| Human resources | 3% |
| Records | 2% |
| Financial statements | 2% |
Success stories: Transforming workflow automation with DocuWare
IDP eliminates manual tasks at a personnel management firm
Piening Personal, a large personnel services provider, manages hourly records for thousands of employees across over 80 locations. The company receives time-tracking reports in various formats due to the differing systems each client uses; some modern, others outdated. Before adopting DocuWare, this information required manual entry or complex rules-based scripts to be imported into their accounting software. Any format changes meant adjusting templates, and odd layouts needed correction.
DocuWare IDP replaced manual work with AI-powered data extraction, flagging only the most difficult cases or those with low confidence scores for review. Now, these routine imports run automatically, reducing peak-period pressure significantly. The IT team is planning to expand IDP to handle handwritten work documents from smaller companies without electronic systems, and processing paper vacation requests.
DocuWare IDP automates processes at an innovative start-up
Sport Auto Plus is an innovative start-up that offers a "car subscription" to people who love driving and motor sports. Customers enjoy a worry-free car experience for one or two years, avoiding leasing contract hassles. With experience from a similar company, the founders take advantage of digital opportunities for a competitive edge and aim to enhance customer communication. The company aims for rapid growth: by July 2024, their website was active, and within six months, over 3,000 orders were received. With plans to expand from 1,000 to over 4,000 cars by the end of 2025, digitalizing processes is crucial.
From the start, DocuWare Cloud incorporated intelligent document processing and an online portal for key documents, essential for customer-facing processes like traffic violation management. The IDP aids in digital tasks, such as invoice processing, by extracting and classifying documents.
Handling minor traffic offences is challenging due to varied notification forms across states. The software extracts offence details, compares them with HubSpot CRM to identify customers, and forwards info to authorities while notifying customers.
Digital invoice processing was live within two months of launch, using a structured approval process and integration with DATEV financial software. While invoices are mostly received digitally, traffic violation notices are scanned and digitized. The IDP achieves a 90% recognition rate for scanned document data.
The digital archive is central to business processes, storing all relevant documents, accessible to customers via a portal, and soon through an app, ensuring transparency and efficiency.
Why intelligent document processing is essential for SMBs
The size and flexibility of small and medium-sized businesses (SMBs) can mean they are more likely to be affected by market instability. Economic pressures like inflation, supply chain issues, and talent shortages make these challenges even tougher, leaving many businesses trying to balance growth with budgetary limitations.
IDP benefits
Replace manual tasks with automated processes: So, SMBs save time, and make better use of their staff’s capabilities. With IDP, your company can maximize productivity without hiring new employees. Less dependency on new hires reduces the costs and effort of recruitment, training, and onboarding. Accurate Data Extraction: Once classification is done, AI algorithms extract key details like text, numbers, and even pictures or signatures. Validation is carried out using techniques like fuzzy logic, rules, and scripts. Human-in-the-loop (HITL) processes improve data accuracy and quickly adjust machine learning models. Easy integration with software in your technology infrastructure: This includes accounting, HR, and CRM software, along with the Microsoft Office suite, SharePoint, and other business tools. Boost customer satisfaction: Businesses that depend on slow, manual processes and don't protect data privacy risk losing customer loyalty. For example, delays in processing insurance or mortgage applications might drive customers to competitors using modern technology. Meet regulatory standards and tighten security: IDP ensures data security and privacy through encryption and stringent controls, complying with standards like HIPAA and GDPR. It prevents unauthorized document access and creates audit trails for regulatory compliance. Gain a competitive edge: With lower expenses tied to manual document handling, resources can be redirected to new initiatives and strategic projects. Future outlook and the way forward
The IDP market is undergoing a significant transformation driven by advancements in AI technology that are challenging traditional practices. Companies that adapt to this change while managing risks will have a competitive advantage in terms of efficiency and customer satisfaction.
Survey results show that managing unstructured data, where IDP excels, is a critical part of AI success. The findings emphasize the need for a comprehensive approach to workflow automation and process mining to find AI improvement opportunities. Additionally, an analysis of the results points out the ongoing necessity to modernize paper-based processes.
