[Accelerating DX in Manufacturing and Food Industries] Entering the Era Where AI 'Learns and Gets Smarter on its Own.' Official Release of 'Knowfa Order Receiving AI Agent V1' Equipped with Autonomous Accuracy Improvement Agent on April 9
UZOUC SYSTEM will officially release 'Knowfa Order Receiving AI Agent V1' on April 9. It features an AI that autonomously learns from human corrections to automate complex order processing.
📋 Article Processing Timeline
- 📰 Published: April 10, 2026 at 02:40
- 🔍 Collected: April 10, 2026 at 09:01 (6h 20m after Published)
- 🤖 AI Analyzed: April 20, 2026 at 08:10 (239h 8m after Collected)
UZOUC SYSTEM CO.,LTD. (Headquarters: Chuo-ku, Osaka; CEO: Naohiro Onoshima) will officially release V1 (the official version) of 'Knowfa Order Receiving AI Agent,' a service where AI replicates the human judgment process in order processing operations, on April 9, 2026. The biggest feature of V1 is the inclusion of the 'Autonomous Accuracy Improvement Agent,' which instantly learns from corrections made by the operator and autonomously improves its judgment logic. By providing just the minimum necessary know-how, the AI autonomously repeats verifications and improves its accuracy, just like a human.
Background: Data Entry Tasks Requiring 'Human Judgment' Unsolved for Decades
In the manufacturing and food industries, purchase orders received via fax or email vary greatly in notation from company to company, and until now, it has been common to rely on the 'intuition and experience' of veteran staff for data entry. It takes about a year for an order processing staff member to become fully proficient, and the drain of business know-how due to leaves of absence or retirement is a serious issue that threatens business continuity.
Our company has been promoting the automation of routine tasks with RPA for over 20 years, boasting an implementation track record of over 1,300 companies. However, non-routine tasks that 'require different handling for each customer' or 'require human judgment' were difficult to automate. To address this issue, we released a beta version in June 2025, gathered field feedback from about 50 companies, and evolved it into V1.
V1's Biggest Evolution: 'Autonomous Accuracy Improvement Agent' and 'Learning Function'
The biggest evolution point from the beta version to V1 is the implementation of a mechanism where the AI autonomously continues to improve its accuracy.
Evolution of V1 (Autonomous Accuracy Improvement Agent)
- AI autonomously conducts verification just by being presented with minimum know-how
- Instantly learns from correction results and automatically brushes up the judgment logic
- Significantly improves the speed from implementation to full-scale operation
What is the Autonomous Accuracy Improvement Agent?
The 'Autonomous Accuracy Improvement Agent' equipped in Knowfa V1 is a function that replicates the human learning and growth process using AI. Just as new employees acquire business skills through OJT, Knowfa deepens its understanding of the business through corrective feedback on site and autonomously improves its judgment accuracy.
- AI autonomously executes a verification cycle based on business know-how
- Instantly learns from the staff's correction data to make correct judgments from the next time (learning function)
- Realizes 'autonomous accuracy improvement' that continuously brushes up judgment criteria
- The scope of autonomous improvement will be expanded sequentially in future updates
How the Learning Function Changes Everything 'From Implementation to Operation'
Conventional AI-OCR only went as far as character recognition and exact match master searches, and could not handle business-level judgments. Knowfa V1 realizes a system where the AI gets smarter the more it is used through a three-layer structure: judgment based on business know-how, automatic referencing of past order data, and instant learning of correction contents.
Background: Data Entry Tasks Requiring 'Human Judgment' Unsolved for Decades
In the manufacturing and food industries, purchase orders received via fax or email vary greatly in notation from company to company, and until now, it has been common to rely on the 'intuition and experience' of veteran staff for data entry. It takes about a year for an order processing staff member to become fully proficient, and the drain of business know-how due to leaves of absence or retirement is a serious issue that threatens business continuity.
Our company has been promoting the automation of routine tasks with RPA for over 20 years, boasting an implementation track record of over 1,300 companies. However, non-routine tasks that 'require different handling for each customer' or 'require human judgment' were difficult to automate. To address this issue, we released a beta version in June 2025, gathered field feedback from about 50 companies, and evolved it into V1.
V1's Biggest Evolution: 'Autonomous Accuracy Improvement Agent' and 'Learning Function'
The biggest evolution point from the beta version to V1 is the implementation of a mechanism where the AI autonomously continues to improve its accuracy.
Evolution of V1 (Autonomous Accuracy Improvement Agent)
- AI autonomously conducts verification just by being presented with minimum know-how
- Instantly learns from correction results and automatically brushes up the judgment logic
- Significantly improves the speed from implementation to full-scale operation
What is the Autonomous Accuracy Improvement Agent?
The 'Autonomous Accuracy Improvement Agent' equipped in Knowfa V1 is a function that replicates the human learning and growth process using AI. Just as new employees acquire business skills through OJT, Knowfa deepens its understanding of the business through corrective feedback on site and autonomously improves its judgment accuracy.
- AI autonomously executes a verification cycle based on business know-how
- Instantly learns from the staff's correction data to make correct judgments from the next time (learning function)
- Realizes 'autonomous accuracy improvement' that continuously brushes up judgment criteria
- The scope of autonomous improvement will be expanded sequentially in future updates
How the Learning Function Changes Everything 'From Implementation to Operation'
Conventional AI-OCR only went as far as character recognition and exact match master searches, and could not handle business-level judgments. Knowfa V1 realizes a system where the AI gets smarter the more it is used through a three-layer structure: judgment based on business know-how, automatic referencing of past order data, and instant learning of correction contents.