Expert "discerning" skills to AI. Streamlining tire development with annotation support

Interview Cooperation: Takai, Chief, Mobility Exploration Business Development Promotion Section 3, Exploration Business Development Promotion Department 2

Bridgestone Corporation Overview

・Established: 1931

・Capital: 126.354 billion yen (as of December 31, 2024)

・Business Activities: Premium tire business, solutions business, chemical products and diversification business, exploration business

Services Utilized

・Annotation Task Support Service

Project Background and Challenges

――HS: Could you tell us the background behind requesting annotation from Human Science this time (trying to outsource annotation)?

Mr. Takei: I am a tire engineer, and there is sensory evaluation information that only engineers who have closely observed the actual objects on site can notice, but it was difficult to convert this into data in terms of accuracy and labor.
By using AI to quantify and numerically express this without taking much time, we can perform various analyses, and as a result, we are working on initiatives that can be reflected in development and proposals for how to use tires.

Therefore, a large amount of data is required, but even if we were to handle it in-house, there are currently not many people with annotation skills, and due to the difficulty and time constraints, the background was that we decided to outsource and have annotation professionals process a large volume of data.

Reasons for Choosing Human Science

――HS: When you decided to outsource, I understand you compared several companies. Could you tell us why you ultimately chose our company?

Mr. Takei: I had interviews with multiple companies, but I felt that Human Science’s explanations were very simple and trustworthy.
There were no exaggerated sales pitches, and they honestly communicated what they could and couldn’t do, which made me feel that the process would proceed smoothly. I remember thinking that.

――HS: I believe you compared us with other companies in terms of estimates and such. Our company was probably not the cheapest, but could you tell us how you considered and decided on us?

Mr. Takei: This time, there were many tasks with very high difficulty, and it was necessary to adjust the specifications each time while progressing with the annotation.
For such tasks, from the consultation stage, I was confident that they could be the most flexible, and I felt that Human Science would be able to achieve the desired results within the budget and required delivery time.

Actual Interactions and Progress

――HS: Could you share any memorable impressions from the actual interactions?

Mr. Takei: Starting with the weekly regular meetings, the exchanges via the Excel Q&A sheet shared on the cloud were actually simple and very easy to manage.
They carefully confirmed with us by asking, "Is this interpretation correct?" regarding what we communicated, and conversely, they left comments on points they were unsure about.
I believe there were multiple people in charge, and from the content of the comments, I could see differences in personality, and it was clear they were trying to resolve issues within the team as much as possible.
There was a sense of trust that if anything came up, "writing it in the Excel Q&A sheet would get the message across."

Significant Improvement in Work Efficiency through Utilization of Annotation Data

――HS: After utilizing the delivered data, were there any moments when you felt the impact or effectiveness within your company?

Mr. Takei: The annotation data you delivered has already begun to be used for model development within our company. Analyses that were previously handled manually have been automated, greatly improving development efficiency.
Tasks that used to require a lot of time and effort are now progressing smoothly, and the staff on site have responded very positively. We even received happy comments saying it feels like we can perform analyses as if by magic.

Evaluation of Human Science Co., Ltd.

――HS: Looking back on the entire project, could you tell us any points for improvement or things you felt were good about requesting Human Science's services?

Mr. Takei: First, as a point for improvement, which is also our issue, I wish we could have secured a little more preparation time before starting. You worked flexibly despite the limited time and the difficulty involved, so going forward, I would like to make requests with more leeway.
Also, I thought that if more people with engineering knowledge joined the Human Science team, there might be more concrete advice provided.

On the other hand, this may be just my personal impression, but my honest feeling was that it was "enjoyable."
When I first made contact with Human Science, I thought they were "an annotation company." Later, after looking into it, I learned that they handle a wide range of services such as manuals and e-learning, and I came to understand that Human Science’s careful and flexible approach is likely a strength cultivated through their experience in education and manual production.

Future Prospects

――HS: Please tell us about your future prospects and any initiatives you would like to expand.

Mr. Takei: Yes. I truly feel that the results of this annotation are a significant asset.
In the AI industry, it is said that "data is everything," and organizations that have well-organized training data with all the necessary information are indeed strong. Well-organized training data like this has the potential to be applied to various models.

To make use of such assets, it is ideal for personnel who have a good understanding of the field to be able to create models with just a little skill improvement. We want to create an environment where not just data scientists, but people with business knowledge can thrive.
In the future, while nurturing personnel who can perform annotation in-house, we aim to establish a flexible system where we can request support from external partners like Human Science when resources are insufficient.

――HS: Thank you very much. We hope to continue supporting you to be of help in the future.

From Human Science Representatives

The project with Bridgestone was a highly challenging initiative to translate the "engineer's intuition" into data.
We are pleased to have been able to aim for a high level of completion together, valuing each and every interaction.
Going forward, we will continue to contribute to our clients' businesses as a bridge between technology and the field.