DATAFLUCT and NX One Bishi Archives Enhance AI Document Management with Azure OpenAI

Introduction

In a significant advancement for the field of document management, DATAFLUCT has collaborated with NX One Bishi Archives to enhance their AI document management feature within the WAN-Sign service. This update incorporates Microsoft Azure's OpenAI technology to achieve an impressive data extraction accuracy of 95.4%. The development highlights how innovative technology can streamline complicated processes such as contract management, thus promoting digital transformation (DX) in corporate environments.

Background: Challenges in Traditional Contract Management

Contract management is essential for any corporate function; however, the manual extraction of crucial information like vendors' names, contract amounts, and expiration dates from a vast array of paper and PDF documents has been a significant barrier to advancing digitalization efforts. To address this challenge, NX One Bishi Archives introduced the AI document management feature of WAN-Sign in October 2024. By leveraging the capabilities of the Airlake platform, this feature enabled automated extraction of necessary items through machine learning technology.

Despite the progress, earlier machine learning models had limitations rooted in their inability to accurately extract information due to the varied formats and expressions found in contracts. This led to inconsistencies and inefficiencies in contract management processes.

Technical Evolution of the AI Document Management Feature

With the recent update, the AI engine responsible for information extraction has undergone a transformation from traditional machine learning models to utilizing Azure OpenAI. This state-of-the-art technology enables the recognition of contextual nuances, going beyond simple pattern recognition to a deeper understanding of language.

Moreover, a hybrid approach has been adopted, allowing both contract image data and pre-processed text data via AI-OCR to be fed simultaneously to the language model. This ensures that subtle textual details and variations in expressions are accounted for, which were previously difficult to capture, greatly enhancing the extraction process’s accuracy.

The seamless integration of data processing through the Airlake platform allows for the effective collaboration of AI-OCR, Azure OpenAI, and structured data extraction of key contract components. This system can now deliver highly accurate structured data concerning essential items such as counterpart names, contract amounts, contract dates, expiration dates, renewal conditions, and cancellation deadlines.

Commentary from NX One Bishi Archives

From the perspective of NX One Bishi Archives, the update has led to notable improvements in processing time and precision, eliciting positive feedback from users. The integration of the Airlake platform has simplified the evaluation process for precision and updates, demonstrating a commitment to enhancing customer workflows and the digitalization of information management with advanced AI technology.

Driving Document Data DX with the Airlake Platform

The goal of the Airlake platform goes beyond contract management. It aims to transform unutilized information assets, such as documents and specifications scattered throughout organizations, into structured data visible through AI technology. By accepting various data formats centrally and allowing AI-driven organization and structuring, this platform plays a pivotal role in enhancing corporate competitiveness.

Case Studies:
1. Design Document Knowledge Management: A major construction company using the Airlake platform's AI-OCR functionality has successfully consolidated disparate data such as drawings, designs, and estimates into a single knowledge source, thereby significantly reducing design time and fostering internal knowledge sharing.

2. Data-Driven Operations in Infrastructure Management: An infrastructure company faced challenges in utilizing departmental data effectively. By centralizing this data within the Airlake platform, they created a dashboard for visualizing equipment status, thus facilitating data-driven operational decisions in response to staffing reductions and aging infrastructure challenges.

Conclusion and Future Directions

Through the Airlake platform, DATAFLUCT continues to promote digital transformation efforts for organizations grappling with complex document processing and data extraction challenges. By improving capabilities to manage contracts, the platform ultimately aims to facilitate comprehensive data-driven decision-making processes across various sectors.

For more information about WAN-Sign AI document management features, visit WAN-Sign's official site and explore how DATAFLUCT strives to enable sustainable decision-making with effective data utilization.