Along with ChatGPT, RAG, LLMs and Prompt Engineering, another key AI term has come along that you should know: Intelligent Document Processing (IDP)—software that uses AI to teach computers to read documents and extract data from them.
IDP is often described as the next generation of document capture or OCR (Optical Character Recognition) technology. While OCR has been around for a while as a way to get computers to read documents and take out and store and forward as needed the informational content we want from them, IDP takes this further by using AI to understand and act on the data more intelligently.
But I think it’s pretty reasonable to ask as a content professional or information specialist if IDP is already mature enough to rely on, or still finding its feet? The answer, perhaps surprisingly, is overwhelmingly “yes.” In fact, new data proves IDP has emerged as a hugely promising technology, evolving rapidly and uncovering new ways to help us.
Surprisingly, its rise is coinciding with a renewed demand for something rather old-fashioned: paper. Whatever the significance of this combination, the findings show that the IDP market is in the midst of a transformative phase, driven by rapid AI adoption, a surge in new document extraction use cases, and rapidly evolving buyer behavior.
A deep dive on IDP
Let’s explore what’s driving this in more detail. We recently sponsored an AIIM (the Association for Intelligent Information Management) study which took a major look into the real state of IDP in the shape of an in-depth survey of 600 US and European OCR/IDP users.
To qualify for analysis, a site had to be an enterprise generating $10M+ in turnover and employing 500 or more people, spanning sectors from public services to energy, financial services, and manufacturing. The findings were striking: 78% of companies are already operational with AI through IDP solutions, often delivered via the cloud as a service from Amazon or Microsoft.
Interestingly, the vast majority (66%) of IDP deployments are new systems replacing legacy ones. Even more striking, much of this new IDP functionality is being addressed on new problems that yesterday’s OCR could not have managed, like scanning and ingestion of licenses and permits, claims forms, new customer onboarding, and medical records.
It may come as a surprise to some that paper is actually helping fuel this growth, with 61% of users saying it’s still a big component of their new IDP processes, and its use may actually increase. So how should we be interpreting these trend lines?
Future-proof, AI-enabled IDP
First off, the reality in late 2025 is that most existing document extraction solutions may not reflect the advanced AI-powered systems that define today’s cutting-edge IDP. While IDP is becoming widespread, our experience suggests that most organizations are still in the enthusiastic, but early-to-intermediate, stages of AI maturity.
Still, the fact that 68% of new IDP projects are replacing existing systems shows that this may not be true for very much longer. Widespread dissatisfaction with current solutions allied to rapid technological evolution means many first-generation platforms failed to meet expectations in terms of flexibility, scalability, or accuracy.
As a result, there’s a growing demand for future-proof, AI-boosted document extraction platforms that enable organizations to integrate evolving AI tools into their document workflows without constant rip-and-replace cycles.
Expansion beyond legacy OCR
Brands also want a lot more than just invoice processing—a high-value but repetitive task that was ideal for early automation. AI has expanded the capabilities of IDP far beyond this, so IDP is moving into customer-facing and compliance-heavy workflows. These are business problems that demand deeper document understanding, integration of structured and unstructured data, and strong audit trails for regulatory compliance.
For example, a KYC flow might involve parsing ID cards with images and text, fraud detection, and archiving in a compliance-ready format. Scaling this with traditional OCR would be extremely challenging; AI-powered content automation provides the speed and intelligence needed to make such applications feasible.
This shift also suggests that IDP is no longer a niche solution but is becoming an essential part of enterprise operations, increasingly integrated with ERP, CRM, and service platforms. That broader integration signals the birth of a new category of, truly “Intelligent Content Automation,” where AI-driven document understanding meets enterprise workflow automation.
I don’t think such Intelligent Content Automation will only come as a commodity service from your friendly cloud hyperscaler. While large cloud providers are playing a growing role by offering infrastructure and native document AI services, these services alone aren’t comprehensive solutions. Business users still need platforms that can orchestrate these disparate tools, integrate them with internal systems, relate them to billions of existing documents, and compose intelligent pipelines tailored to business needs. In practice, this necessitates agile integrators and orchestrators that combine hyperscaler services and proprietary tools to deliver flexible and business-centric IDP solutions.
Paper? Far from dead
Which brings us to paper. Despite immense digital transformation efforts and campaigns like AIIM’s “World Paper Free Day” every November, 61% of IDP processes still involve paper. Notably, 48% of respondents actually expect their use of paper to grow.
This shouldn’t be too surprising. The resilience of use of paper is often tied to regional habits (e.g., the ongoing insistence on paper checks in the US) and regulatory requirements (e.g., legal mandates to retain physical copies). That said, legislation in Europe such as mandates for e-invoicing, is gradually reducing reliance on paper.
Personally, I think paper will gradually meet us halfway, becoming increasingly digitally enhanced, with barcodes and metadata that enable more efficient automated processing.
IDP strategies
The IDP market is at an inflection point, with AI acting as a catalyst for growth and a disruptor of legacy norms. Let’s accept that IDP is becoming central to many new business processes and deserves serious attention, and so you should prioritize IDP solutions for high-impact, customer-facing use cases.
To achieve this, conduct rigorous evaluations of potential vendors, focusing on their data security, GenAI readiness, ease of integration, and their end-to-end capabilities for not just document understanding but also automation and the ability to understand the context of how a decoded document relates to other documents. Would I like to see less paper as part of this transformation? Sure, but I think IDP will actually soon be so effective that, like fossil fuel, paper may soon be at its peak.
The author (Dr John Bates) is CEO of SER Group, a global leader in ECM
