GenAI breathes new life into IDP but will eventually replace it

Enterprises continue investing in intelligent document processing (IDP) as they dig deep to find productivity gains by reducing reliance on paper and improving their data infrastructure. But is it money well spent? Are there other more advanced alternatives to consider?

Think beyond IDP to move quickly out of the unstructured data trap

The IDP space comprises a compendium of interwoven technologies, including OCR, NLP, document extraction and classification, computer vision, ML, AI, and even RPA (some say). Whatever you consider IDP to be, the advancements in generative artificial intelligence (GenAI) have opened a whole new set of opportunities across the enterprise to create and manage content and data. Enterprises need to rethink how to ingest unstructured data in more meaningful ways and consider whether, with GenAI, they still need to consider anything “unstructured” at all.

GenAI accelerates IDP while creating a path to pragmatic digitalization

If artificial intelligence (AI) can make sense of almost anything we see, read, or hear using large language models (LLMs), then surely many of the barriers to transformation that led to the creation of the IDP space in the first place must come down. Thanks to Gen AI, we can now ingest vast amounts of data with more context and meaning. When we combine this “unstructured” data with all the other aspects of the workflow, the customer (context and intent), and all connections to the “data,” we can see a real change in how enterprises cut customer service costs and actually serve them a whole lot better.

In a true Generative Enterprise™, GenAI programs must work in harmony with IDP

When we look at the broader picture of the future Generative Enterprise™, we can see it will become less about bolt-on AI technologies and more about AI-connected ways of doing business. As AI becomes better at interpreting complex data, bolt-on AI should be your first choice only if your needs are outpacing integrated AI innovation. You can use NLP to read and understand emails, but what if you combine that knowledge with advanced AI to ingest and process attachments, existing contracts, and live customer data?

Using retrieval augmented generation (RAG), for example, we could then go from automatically solving 75% of interactions to solving 99%. As we connect AI with each core element across enterprise workflows, we can also ensure we are ready to add human oversight only where it’s really needed. We can also start to be proactive instead of reactive as we use AI to detect business goals that might not be met long before we don’t actually meet them! We need to prepare for all of this. We need to invest in the right interconnected (or connectable) AI and adopt more agile business modes.

IDP covers silos of legacy tasks, but AI done right rejuvenates the legacy

Some have been strong proponents of the paperless office for many moons, but it’s still many moons away, although each year, we see many digital native enterprises that don’t even bother with paper at all.

But with AI, does it even matter? Analog paper, for example, plus AI is digital in all but name! The line between structured and unstructured data blurs. We are already using AI to read and ingest convoluted forms the doctor’s office still requires of us all (in the USA, anyway). We can now read and make sense of large, complex purchase orders and invoices that would previously have kept many a human awake at night. We can decipher and present summaries, contract highlights, and breaches and even present legal arguments from volumes of documented precedents that are hard to know exist, let alone find. This is no small feat, but it’s here now, and it’s only getting more advanced with every passing week.

For superior outcomes, infuse AI holistically across the enterprise, not in silos of Band-Aids

Many IDP vendors recently announced GenAI solutions to complement their product sets. UiPath and Automation Anywhere, for example, recently announced their AI products and AI-related strategies. Time will soon tell whether their strategies are more of the same tactical bolt-ons riding the AI wave or if they reflect broader thinking about AI across the enterprise to complement the true meaning of the Generative Enterprise™. As more software platform vendors like Appian and Pega, for example, include IDP technologies in their stacks, it provides yet more confirmation of this being where they and their enterprise clients are headed in this narrative. See our recent IDP Horizons report, HFS Horizons: Intelligent Document Processing Products, 2023.

Before you turn to IDP—or any standalone technology—for the less structured and less connected parts of your business, consider first this Generative Enterprise™ narrative, which demonstrates what you need to think to take your business safely into the next era of transformation.

The Bottom Line: AI must not live in silos. Instead, it should permeate all that you do.

This approach will still support an enterprise strategy with IDP technologies where they apply, which is not everywhere. You will buy yourself a little more time as this AI strategy will not go to waste. As you phase out IDP—and you eventually will—the business areas where you have holistically applied AI will likely be reusable and infusible into all that you do next.