#Japan #Tokyo #invox #AI OCR #Infomart

Introducing the New 'Reading AI Agent' in Invox's Order AI-OCR

In April 2026, a significant update to Invox's Order AI-OCR service will introduce the groundbreaking 'Reading AI Agent'. This innovative feature promises to revolutionize order processing by allowing AI to learn and interpret user instructions, resulting in highly precise data extraction, thereby drastically reducing the time and labor involved in order processing.

Background of the Feature Release

The digital transformation of back-office operations in companies is accelerating. However, challenges remain, especially when dealing with order forms like purchase orders from various vendors. These forms often exhibit diverse layouts and formats, which lack standardization, presenting considerable hurdles for traditional AI-OCR technologies.

This is particularly evident in the food service industry, where order forms are frequently complicated. These forms may specify multiple delivery dates or locations for the same product, making it difficult for conventional AI-OCR systems to accurately recognize and process the data. As a result, order processing often demands manual oversight, with personnel spending time verifying and inputting data.

To address this, Invox is launching the 'Reading AI Agent', designed to facilitate greater efficiency in order processing. Utilizing advanced AI technology, this feature aims to enable accurate, intent-driven data extraction, drastically minimizing the time and complexity involved in handling orders.

Feature Overview

Reading Complex Orders

With the 'Reading AI Agent', users will now be able to directly instruct the AI on the rules for reading orders and converting the data. This enhancement allows for the handling of complicated orders, such as:

• - Orders with multiple delivery dates or locations for the same item.
• - Orders with items arranged in irregular patterns.

This advance eliminates the previous need for specifying reading coordinates for each form, leading to significant improvements in accuracy and speed in data entry tasks. Moreover, this feature leverages large language models (LLMs), allowing users to set intricate processing rules in everyday language. For instance, one can easily instruct the AI to retain a previous delivery date if the new entry is blank.

Efficient Sorting and Review

The AI will enable sorting of received orders based on delivery locations and dates for easy review. This means that users can quickly manage detailed order information without the burden of manually verifying every entry, particularly for larger datasets.

As a commitment to continuous improvement, Invox aims to expand the functionality tailored to customer needs, focusing on achieving 100% digitalization of order and invoicing processes within the food industry while enhancing digital efficiency across corporate back-office operations.

Customer Feedback from Early Users

Prior to the official launch, the 'Reading AI Agent' was trialed by Oka Industries, receiving acclaim for its innovative features.

Challenges Encountered Before Implementation

Oka Industries faced issues with unique order forms from clients, as slight misalignments or scale changes led to frequent inaccuracies in reading. Each instance necessitated correcting the layout, a process that could take up to five minutes per document. Additionally, the food service sector's characteristic multiple delivery dates led to further manual processing time.

Results After Implementation

Post-implementation, the 'Reading AI Agent' offered a game-changing improvement. The AI can now accurately gauge the content without frequent layout adjustments, virtually eliminating reconfiguration delays. As a result, the processing time for each order was reduced by 60-70%.

Stable reading accuracy now allows for a broader range of orders to be processed through the AI-OCR system, enhancing the company's ability to shift from labor-intensive tasks to value-added analysis and planning, thus driving workplace reform.

Official Comments from Infomart

Saito Fumihiko, Corporate Executive Officer at Infomart, stressed the organization's commitment to achieving '100% Digitalization of Order and Invoicing' in the food industry. Despite previously implementing various digital solutions, the challenge of complex order formats remained a notable barrier, particularly in food services. The latest 'Reading AI Agent' is a revolutionary advancement that harnesses LLM technology to interpret human intent, enabling the efficient processing of orders and supporting the digital transformation of the broader food sector.

Upcoming Seminar

In line with the launch of this new feature, a demonstration seminar will occur on April 14, 2026, at 14:00. Attendees will experience this next-generation AI-OCR's capabilities, which promises to significantly cut processing time and effort.

Details of the Seminar

• - Date: April 14, 2026
• - Time: 14:00 - 15:00
• - Fee: Free
• - Format: Online (via Zoom)
• - Registration URL: Register Here (Registration deadline: April 14, 2026, at 13:00)

Contact for More Information

For inquiries regarding this feature, find further service details here.

About Infomart

Founded in 1998, Infomart has been at the forefront of cloud services aimed at enhancing business efficiency in inter-company transactions like invoicing and order placements. Its flagship platform boasts over 1.25 million users, managing transactions amounting to more than 71 trillion yen annually. Infomart's dedication to streamlining B2B electronic trading continues to evolve, reinforcing its commitment to innovative business solutions.