Yokoyama

We believe it is essential to continuously support and improve the quality of proposals that insurance agencies provide to their customers. Some agents already grasp the financial status of their client companies through financial statements and maintain a detailed understanding of factors like employee counts through frequent visits. However, it requires significant effort for individual sales personnel to track such information. We aim to support the sales activities of these agencies by providing them with high-quality, reliable information that serves as a catalyst for proposing the necessary insurance to their customers.

Yokoyama

We began considering the integration of corporate data for our business clients into "MS1 Brain," the first sales support system for insurance agencies in the non-life insurance industry to feature AI, back in 2018. I clearly remember that at the time, uSonar's corporate database, LBC, significantly outperformed other companies in terms of matching rates.

Yokoyama

At the time, we faced challenges regarding news distribution to our agencies. When an agency visits a client company, understanding the company's trends and industry information in advance is fundamental to sales activities. Therefore, we were distributing news from major newspapers to agencies through MS1 Brain, but the annual cost became a burden. Furthermore, since the distribution was based on search results using keywords related to specific industries, we faced the challenge that we could not provide information effectively linked to individual companies.

Yokoyama

The catalyst was the implementation of uSonar for our CRM development project for employees. During the data accuracy verification process, it demonstrated an extremely high match rate. Furthermore, when we had them perform data cleansing for our company, they identified and resolved duplicate data in no time.
At that time, I noticed that the corporate attribute information known as "Story"* has been significantly enhanced compared to 2018. This includes information such as whether a company is certified as a Health & Productivity Management Organization, has obtained ISO certification, or is a member of a Chamber of Commerce. While this information can be found through research, manually checking each company would require an enormous amount of time and effort. The fact that we can easily and comprehensively obtain this information and provide it to agencies through MS1 Brain is of great value in supporting the proposal-based sales activities of our agencies.

Yokoyama

The freshness of the information was also a deciding factor. The corporate database we were using previously had a monthly update frequency, and we felt there were challenges in reflecting the latest information. uSonar's LBC allows for daily acquisition of the latest data, enabling agencies to conduct proposal activities based on fresh information. Additionally, being able to utilize the latest data for analysis was also a crucial point.

Yokoyama

General corporate information services primarily focus on data from companies that cooperate with surveys and questionnaires, which inevitably leads to a bias toward companies that are easily accessible. However, the clients that agencies interact with daily range from publicly traded companies to small and medium-sized enterprises and family-owned businesses, and information on such companies is exactly what is in demand. LBC's highly comprehensive corporate information covers that gap. This was a crucial point for us.

Furthermore, we were pleased to discover that by implementing uSonar, we could distribute various news items linked to individual companies, which provided a prospect for significantly reducing the news distribution costs that had been a challenge. By also introducing SideSonar into our CRM for employees, we resolved concerns regarding discrepancies in the corporate information referenced by employees and agencies. Determining that multiple issues could be solved at once, we decided to implement uSonar during our system renewal in 2025.

Yokoyama

We have integrated it into the customer information screen of "MS1 Brain" so that story information, news, and analysis scores can be viewed in a list. You can think of the analysis score as an indicator of product recommendation suitability. We create machine learning models to calculate these scores, but this requires a massive amount of data. We utilize LBC data, including story attribute information, for that purpose.

Yokoyama

For example, we aggregate data into "features" in formats such as "whether they are a member of the Chamber of Commerce" or "whether they have ISO certification." For annual certification information like "Certified Health & Productivity Management Outstanding Organizations," we organize it based on business significance, such as whether to use only the most recent fiscal year or to convert it to indicate a history of certification. It is necessary to change how data is handled depending on the analysis objective, and there is no single correct answer. We organize the data with ingenuity and use it as input data for machine learning.

Yokoyama

When we conducted a comparative verification using an AutoML tool, comparing cases with and without LBC story information (see photo), the results generally showed higher scores when LBC data was included.

Yokoyama

Agents log in to the system we provide for them. For example, when they select a client whose contract is nearing expiration, they can collectively review various information, such as their current insurance contracts, correspondence history, and recommendations for insurance products with high suitability. It is also possible to check company information and news linked via LBC codes.

While the story information includes data they may already know, we believe it is effective for reviewing information before visiting clients, considering proposal content, and uncovering potential needs because it allows them to comprehensively confirm attributes and discover information they were previously unaware of. Although usage frequency varies by agent, many use it on a daily basis.

Yokoyama

As for quantitative benefits, I would point to the reduction in running costs and the improvement in analysis accuracy through machine learning mentioned earlier. However, I believe the true value and intended effect lie in the qualitative aspect: by supporting proposals to clients—which often tend to rely on experience and intuition—with data, we have become able to deliver higher-quality proposals that are appropriate for our clients.

Yokoyama

Moving forward, I would like our employees and agents to utilize it even more on the front lines. For example, I envision our sales staff and agents discussing sales strategies, such as, "Looking at the story, Company A has these characteristics. Therefore, let's recommend this product and service using this proposal method."

We highly value the LBC data content provided by uSonar. As you continue to evolve the quality and quantity of the information provided, we intend to further expand the scope of our utilization.