日本語 
WAIOS Co., Ltd. will collaborate with QuackShift, a startup spun off from Matsuo Research Institute, to realize “semi-automated AI”.
As the first step, we have developed an implementation example focusing on the initial review of drawings before submitting them to a supervisor, making it easier for the person in charge to identify and correct any shortcomings themselves. This will be unveiled for the first time at the Manufacturing DX Exhibition (booth number 27-58) at “Manufacturing World Nagoya,” held from April 8th (Wed) to 10th (Fri), 2026.
WAIOS’s concept of “semi-automatic AI”
“Semi-automated AI” is a concrete manifestation of the WAIOS concept, making it easier to experience in real-world settings.
This system identifies missing information, omissions, or illegible sections while the person in charge is inputting, creating, or reviewing the document, or just before submitting it to their supervisor, and returns it in a format that can be corrected by a human.
Also Read: Kaminashi and Izigen Prioritize Factory Floors Over Slides
In other words, it’s not about “AI doing everything,” but rather “AI that stops and returns before a human can even submit something.” The goal isn’t to have AI make design decisions, but rather to make it easier for the person in charge to notice shortcomings and correct them themselves before submitting to their superior.
In collaboration with QuackShift, a startup spun off from Matsuo Research Institute, we are making “preliminary review before drawing submission” a reality.
WAIOS will collaborate with QuackShift, a startup spun off from Matsuo Research Institute, to bring this semi-automated AI to life.
This time, the main topic is “pre-submission initial review of drawings,” which effectively communicates value in the manufacturing field.
Just before the person in charge submits the drawing to their supervisor, we review it using screenshots, PDFs, and images to identify any missing or illegible information regarding key aspects of the initial prototype, such as the drawing number, revision section, scale, material, and quantity.
The AI will point to the relevant area in the image and respond with short comments such as, “The drawing number field is missing,” “The text in the scale field is illegible,” or “Please check the material field.”
The person in charge can review the content, make corrections, and request further confirmation if necessary.
The expected benefit of this system is that it will make it easier to reduce the number of times applications need to be reviewed or returned for revision.
In addition, it is expected to reduce the psychological burden on the person in charge, such as “I don’t know if this is okay to submit” or “I don’t want to get in trouble for submitting this.”
Especially in manufacturing settings, when it comes to verification processes related to quality and delivery dates, the value lies more in preventing things from proceeding automatically than in simple automation.
Stopping when you’re unsure and returning the information to others ultimately leads to easier implementation and a greater sense of security.
SOURCE: PRTimes
English