Future Proof Solutions: Embracing AI Advancements in Intelligent Process Automation

The heart of the TCG Process mission is helping organizations achieve greater operational efficiency and agility, which is reflected in our priority of delivering future-proof solutions.

The cornerstone of this approach is based on the foundational approach of striving to have the strongest process management and orchestration capabilities in the market, combined with a rich set of pre-built functional components (we call them activities) that bring processes to life. As we find ourselves in the middle of a technology storm around AI, it is an open approach and the ability to simply plug-in new technologies that delivers the best choice for organizations.

In this blog post, I will explain how and why AI choices can be adaptable, and how at TCG Process we are embracing new technologies to maximize customer ROI and expand on new use cases for intelligent document processing (IDP).

Who is going to win the race for the ‘best AI’?

We are in an exciting and yet somewhat unpredictable point in time with AI offerings dominating the headlines and a race to ‘win’. But it’s early days, and there are a lot of things that remain unanswered. When I posed the question to ChatGPT about who might win this race, the response generally aligned with our point of view:
“As an AI, I don't have access to real-time information or the ability to predict the future. Therefore, I cannot accurately determine who will win the race for the ‘best AI’ in the future. The development and progress of AI technology involve numerous factors, including research advancements, technological breakthroughs, and the contributions of various companies, organizations, and researchers. It's a highly competitive and rapidly evolving field, and the outcome is uncertain. We will have to wait and see which advancements and innovations will shape the future of AI.”

It is evident that we are merely at the initial stage of this journey, and that AI will continue to evolve, adding value in ways we haven't even thought about yet. Given this level of unpredictability, how does one make the right choice? Our experience has taught us that this fear is one that we want to mitigate – not needing to make such a choice at all, but rather being able to take advantage of the best of all worlds by seamlessly connecting 3rd party services and complementing them with specialized data and document handling capabilities of its own.

“We are entering the platform era”

I heard this statement from an EY consultant during a recent conference. He elaborated on its meaning, emphasizing the idea of companies integrating a series of interconnected technologies within a unified and coordinated architecture. This approach enables businesses to fulfill their requirements faster and more effectively than ever before. Additionally, it provides the flexibility to complement and/or replace any individual components without compromising the overall platform.

This all sounds rather familiar. At TCG Process, we have been embracing this concept for years through our open process automation and intelligent document processing (IDP) platform, DocProStar. Our built-in orchestration capabilities position us perfectly to leverage the latest and best technologies available, combining both home-grown and publicly available AI services that are driving new use cases and applications. We can easily swap or combine technologies to achieve optimal results without any disruption to the critical processes that our customers rely on, whether they are sourced from the cloud or on-premise.

How we are putting AI to work

AI has almost always been part of our solutions. Historically this has included rules-based symbolic AI, which refers to a category of artificial intelligence that relies on human-created rules and logic to emulate the necessary intelligence. These have played a role in almost all projects that involve document classification and data extraction.

Large language models, or LLMs, and other AI models, represent an exciting opportunity for new use cases and for accelerating project timelines. LLMs extract data differently from traditional data extraction technologies and even AI-based services like Google Vision and Microsoft Cognitive Services. But with Google’s Bard and Microsoft’s investment in OpenAI, we can expect to see Google and Microsoft building these capabilities into their more advanced AI-based services.

When it comes to data extraction, traditional OCR was primarily focused on performing character-level recognition without understanding the context or semantics of the text. LLMs offer remarkable language understanding and can handle a wide range of unstructured textual data, but legacy technologies and specialized services may offer more accurate and efficient results in specific scenarios, particularly when dealing with structured forms, real-time processing and character level accuracy requirements, or when compliance and data quality are the highest priority.

Ultimately, the choice of technology really depends on the specific use case, the nature of the data, the desired level of accuracy, where data can/should be processed and even resource constraints, which is why we provide a future-ready platform that welcomes AI advancements. The openness of our platform means we have already been leveraging more advanced AI-services like Google Vision and Microsoft Cognitive Services for quite some time, and we are now embracing totally new use cases by leveraging LLM’s, generative AI and other ML and AI techniques using activities that seamlessly embed those technologies directly inside of a business process. In my next post I will explore more about various AI models and deployments can best improve IDP outcomes.