TechnoPro, Inc. TechnoPro IT Company provides "Working Voice," a service that allows data entry by voice during inspections, record-keeping, and other tasks in factories. We spoke with Mr. Suda, manager of the System Solutions Division, about the background to the use of voice, the product development process, and why they adopted the AmiVoice SDK.

Issues/background

The inefficiency of taking notes on paper during factory inspections and testing is an issue

Currently, when working on-site in factories and other places, keyboard input is difficult because workers wear gloves.
Therefore, for example, when performing quality control tasks such as inspections and measurements, workers must write down the information on paper while still wearing gloves, and then enter the data later in the office.
This method required the double effort of taking notes and then entering the data later, and also posed issues with the reliability of the data due to mistakes made when writing by hand or typos when entering the data.
To solve both these problems of "poor work efficiency" and "ensuring data reliability," we developed a voice-based input system called "Working Voice."

Key Drivers of the Adoption

High recognition rate of alphanumeric characters and words and usability in noisy environments

Quality control work mainly involves inputting numerical values ​​and product numbers during inspections. For this reason, we adopted AmiVoice SDK, which has high accuracy in recognizing alphanumeric characters and words.
In addition, factories are noisy environments that are difficult for voice recognition, but the system provided by Advanced MediaAmiVoice Front WT01Another big attraction was that by using this microphone in conjunction with the device, it was able to recognize voices even in noisy environments.
One of the major points is that you can easily register words you want to recognize by simply assigning the reading in hiragana to the notation.

Difficulties encountered during development and how they were resolved

Since I had attended the technical training course at the time of implementation, the implementation went without any problems.
However, while the system recognized voices well in quiet environments, the recognition rate decreased as the noise level increased.
I was struggling to figure out how to tune it and other solutions, but I was able to solve the problem by asking Advanced Media for help.

Introduction Results

Increased productivity not only in record-keeping but also in the entire business process

When carrying out inspections or measurements, we used to have to stop work and record the measurements by hand after each measurement, but now we can input the data by voice without stopping work, which has greatly improved work efficiency.
In addition, the process of later converting handwritten notes into data has been eliminated, and data can now be converted instantly from voice messages, reducing input errors and increasing the productivity of the entire business process, including subsequent report creation and data integration with core systems.