WallTech OCR × Amazon AI Model Debut | Achieve 99% Accuracy & 30% Efficiency Improvement

International freight forwarders are resource integrators, coordinating upstream and downstream elements of global logistics. Nearly every company engaged in international trade relies on at least one freight forwarding company as a partner.

As coordinators of various business activities, freight forwarders must process a multitude of documents from each partner around the clock, including packing lists, contracts, invoices, bills of lading, customs declarations, booking orders, health certificates, licenses, certificates of origin, and more.

Document processing in a system requires the digital or physical file to be received, followed by copy-pasting or manual input of relevant fields into the system, with a thorough review for accuracy.

According to WallTech's customer research, a single booking order takes at least 3–5 minutes to complete from copy-pasting and manual entry to the final review.

However, traditional OCR (optical character recognition) technology integrated into most freight forwarding systems lacks the accuracy needed, often falling short of the 99% threshold and requiring a second round of verification, making form recognition time-consuming and labour-intensive.

With numerous associated partners and a wide variety of document templates, each new template requires the system to be individually trained, leading to high onboarding costs for new forms.

To improve the accuracy and efficiency of logistics document recognition, WallTech has developed a new OCR intelligent form recognition application powered by Amazon Web Services (AWS) and Amazon Bedrock’s industry-leading large language model. This enables large-scale form recognition with accuracy rates reaching 99%, enhancing overall user efficiency by 30% and significantly improving customer satisfaction.

From October 31 to November 1, 2024, at the AWS Generative AI Co-Innovation Conference held with Kingsoft Office under the theme “Moving Forward with AI,” WallTech presented its multimodal intelligent OCR recognition solution powered by Claude 3.5 and showcased its application in the international logistics industry for the first time.

At the Generative AI Co-Innovation Conference, the WallTech and AWS teams provided live demonstrations of complex booking documents and various invoices being uploaded and recognized.

Abandoning traditional OCR and data processing methods, WallTech’s solution leverages advanced AI models to evolve from image recognition to image understanding, and from text processing to text comprehension, further enhancing final accuracy.
From an architectural standpoint, the system simplifies the overall structure, moving from dependency on image recognition, NLP, and text processing to relying solely on the interpretive capabilities of a large language model.
The intelligent booking order recognition function supports AI recognition without strict format constraints, compatible with pdf, word, excel, and image files, as well as bulk upload and recognition.
Once recognized, users can generate orders with a single click, automatically populating basic business order information, eliminating the need for manual entry. It also supports multi-line cargo description recognition and bulk order generation.
Additionally, CargoWare enables the automatic assignment of work tasks based on roles. With a set of associated rules, information such as port routes recognized from the documents can be linked to personnel in operations and customer service. Partner information, such as destination port agents, can also be linked, automatically entering the booking details and associated personnel information into the system, thereby achieving automated data entry and work assignment.

AI models are sparking efficiency revolutions across a multitude of industries.

As a pioneer in the international logistics SaaS sector, WallTech will continue to explore more high-quality AI solutions, integrating them with its logistics cloud services to drive the practical application and intelligent upgrade of logistics operations, improve efficiency and quality, and propel the traditional international logistics industry into a new era of generative AI development.

About WallTech

WallTech (China) Co., Ltd, founded in 2011, is an industry-leading developer and operator of cross-border logistics SaaS cloud service platforms. WallTech focuses on international freight forwarding logistics and cross-border e-commerce logistics software cloud services, dedicated to the implementation of SaaS technology in the industry via world leading infrastructure provider such as AWS (Amazon Web Service), and strives to become the most trusted platform for international logistics customers.

WallTech, leveraging Amazon Web Services, has developed an intelligent form recognition application that utilizes the multimodal capabilities of LLMs to improve form recognition accuracy and expedite new form recognition. Simultaneously, an enterprise knowledge base was built to enhance internal efficiency and streamline logistics service applications for external use.

Currently, WallTech relies on a suite of Amazon Bedrock、AWS Lambda、Amazon Aurora、Amazon SageMaker、Amazon OpenSearch Service、Amazon Elastic MapReduce（Amazon EMR）、Amazon Redshift、Amazon Relational Database Service （Amazon RDS）、Amazon Simple Storage Service（Amazon S3）、Amazon Elastic Compute Cloud（Amazon EC2）, among others.

For more information about WallTech's logistics IT solutions and a demonstration of the latest OCR function, simply visit www.etowertech.com and book for a demo.