Elevating Corporate Governance And Stewardship With AI

Generative artificial intelligence (AI) is beginning to reshape how institutional investors approach stewardship and corporate governance. Nippon Life Insurance Company uses AI technology to analyze the quality and impact of stewardship engagement across more than 700 investee companies. By analyzing the quality of dialogue with the corporate investees, the firm is catalyzing meaningful changes in corporate behavior over time.

In a recent conversation, Oliver Wyman partner Yasuyuki (Yash) Sekioka speaks with Tomochika Ishii, CFA, General Manager of Nippon Life’s Equity Investment Department, to explore how AI technology is beneficial to improve the quality of the stewardship dialogue.

INFocus Series

INFocus provides exclusive insights and trends from experts and leaders across the Asia Pacific region, exploring the forces, opportunities, and challenges shaping its future.

Go to series

Yasuyuki (Yash) Sekioka

Hello everyone. I’m Yasuyuki Sekioka, partner at Oliver Wyman, Tokyo office. Today, we are joined by Tomochika Ishii, CFA, general manager of Equity Investment Department, Nippon Life, who will be discussing with us Nippon Life’s stewardship activities. We also hope to discuss Oliver Wyman’s support in those endeavors. With that, Mr. Ishii, thank you for joining us. 

Tomochika Ishii

Thanks for having me. I appreciate this opportunity. I am excited to discuss Nippon Life’s stewardship activities, which have been ongoing for over ten years.

Yash

Thank you, Mr. Ishii, I look forward to it. First, could you tell us what kinds of initiatives you have undertaken in stewardship activities over the past ten years at Nippon Life? Particularly amid environmental changes, I would like to hear how you responded to those changes.  

Tomochika 

Over the last 10 years, in pursuit of economic progress, corporate governance reforms have made a lot of progress. In 2014, we saw the establishment of Stewardship Code for Institutional Investors, and the Corporate Governance Code for companies in the following year. Acting in tandem, these two pillars have consistently advanced governance reform.

In this context, as a responsible institutional investor, we announced in 2014 that we would adopt the Stewardship Code and, while strengthening our framework, have worked to enhance both the quantity and quality of our dialogues. Specifically, in terms of quantity, we currently conduct dialogues with around 700 investee companies across Japan each year, and we believe we are among the most active in Japan. 

The themes covered in these dialogues have also gradually expanded. At first, most revolved around topics of governance and finance, but starting from 2018, we gradually expanded to themes such as the environmental and society as well. In terms of what we aim for, we want to be an institutional investor that is chosen by companies as a valuable dialogue partner. With that in mind, we engage in dialogue with many companies. We accumulate insights and use them to improve the quality of our dialogues. By doing so, we hope to keep evolving our activities going forward.  

Yash

Thank you very much. We now have a good understanding of what Nippon Life has done in stewardship over the past ten years. 

In this project with Oliver Wyman, we analyzed the dialogue meeting minutes from over ten years of ongoing activities in Nippon Life, to determine whether Nippon Life has truly conducted high-quality dialogues and whether they have been properly received by the investee companies, and further based on those dialogues, whether investee companies have actually undergone behavioral change that could be confirmed externally as observable outcomes. We supported you by analyzing and confirming these points.

To make our analysis as objective as possible, we used generative AI. First, we trained AI to understand Nippon Life’s investment philosophy. Then, we trained it to recognize what constitutes a higher-quality dialogue, and we had it read all the dialogue minutes to assess whether those dialogues were of high-quality and were creating impact. When launching our project, where there any particular expectations you had of Oliver Wyman? 

Tomochika 

We chose Oliver Wyman as our partner for three reasons. First, we have worked together with Oliver Wyman on stewardship projects over the past two years, which meant Oliver Wyman already had a deep understanding of our activities and philosophy. 

Secondly, even on a personal level, I really had the feeling that we were able to share the same goals and the level of quality we were aiming for — even before the concept stage of this project. 

Thirdly, the project was truly challenging for us, and we were expecting to hit many roadblocks along the way. Within that context, we expected Oliver Wyman would likely provide appropriate and useful solutions. Those were our expectations, those were the three reasons. 

In this project, you met those expectations. In fact, you exceeded them in two ways, and I would like to expand upon these. The first one, as you mentioned earlier, was the analysis done on the dialogues’ quality. This is where most time was spent. The nature of this endeavor on our end was to instruct generative AI on what constitutes an effective dialogue, and to get it to learn and assimilate accordingly across our 8,830 records so that it could appropriately assess the dialogues’ quality. That was the process. That is why the training and prompts were extremely important.

We started with a 300-record sample and iterated by trial and error. When you do trial and error with generative AI, while it excels at understanding context, it can also interpret our statements too favorably, so during the sample-checking stage, we performed tasks such as excluding such cases. In the end, the prompt’s contents came to roughly the equivalent of ten A4 pages. Also at that time, Oliver Wyman provided very practical solutions in our view. That was the first point.

Regarding the second point, we needed to assess, through generative AI, changes in corporate behavior. The procedure may look simple, but it is not so simple as we make various requests to the companies we invest in. For example, quantitative requests on the financial side or various qualitative requests such as enhanced disclosures. Given that context, the question was: how do we manage this systematically?

So first, we had generative AI to go through the requests from the dialogue minutes — over 10,000 of them— – and have it read all of them reliably, classify them into quantitative or qualitative requests, and pre-load and cross-reference with financial data on the quantitative side and disclosure materials such as securities reports and integrated reports on the qualitative side. By pre-loading these into the system environment, we could match our requests with the system’s data, and have the generative AI assess changes in disclosure documents and financial data. 

We asked your team at Oliver Wyman to perform those tasks, but to be honest, we were half-skeptical whether generative AI could really make such sound judgments. However, you managed to build the system environment admirably and made reliable judgements possible, which truly surprised us. Thank you very much.

Yash

Thank you very much. Through this project, looking back at Nippon Life’s stewardship activities over the past ten years, were there any new insights or discoveries? Could you share these? 

Tomochika 

In this analysis, the proportion of dialogues that were assessed as good in terms of quality and company response increased year by year, which we view positively. However, as expected, there were differences depending on the theme of the dialogues. For example, for dialogues on financial topics, relatively speaking, the proportion rated “High”, “Good”, or “Effective” was higher, whereas for themes like environmental and society, the proportion was relatively lower.

Each of our stewardship officer handles just under 100 investee companies per year. In that context, in order to raise the proportion of effective dialogues going forward, we felt we need to further consider the way in which we should explore our future activities. 

Regarding changes in corporate behavior, the proportion of cases where change occurred has increased year by year. Having said that, this is not to claim that the effect is solely due to our dialogues. Companies are influenced by many other factors, for example policy and requests from other institutional investors. Change can also arise from the company’s own corporate initiative. Among all these influencing factors, we see general alignment between our requests and the direction that companies take in terms of behavioral change. We see consistency in these results. We also found that when we conducted effective dialogues and received positive responses from companies, the proportion of change in corporate behavior was relatively higher. This confirmed that our activities were going in the right direction. 

Yash

Mr. Ishii, thank you for your valuable comments. Lastly, could you share Nippon Life’s view of stewardship, as well as your own vision, for future activities and initiatives? 

Tomochika

We believe that companies themselves must recognize the need for improvement and change of their own accord in order to improve corporate value in a sustainable manner. That is what we hope for. Recent discussions on Japan’s corporate governance reform focus on improving companies’ income sustainably and to that end, we will continue to engage with our investee companies. We believe many of our requests will take about 3 to 5 years, or longer, to achieve our requirements. With that in mind, we intend as a long-term-oriented institutional investor, to continue making requests to companies and support them in their initiatives. That is what we would like to do going forward. 

Yash

That was very clear. Mr. Ishii, thank you very much for today. 

Thank you for watching. For more information on this topic, please visit Oliver Wyman’s LinkedIn page and make sure to follow us for more updates.

Generative artificial intelligence (AI) is beginning to reshape how institutional investors approach stewardship and corporate governance. Nippon Life Insurance Company uses AI technology to analyze the quality and impact of stewardship engagement across more than 700 investee companies. By analyzing the quality of dialogue with the corporate investees, the firm is catalyzing meaningful changes in corporate behavior over time.

In a recent conversation, Oliver Wyman partner Yasuyuki (Yash) Sekioka speaks with Tomochika Ishii, CFA, General Manager of Nippon Life’s Equity Investment Department, to explore how AI technology is beneficial to improve the quality of the stewardship dialogue.

INFocus Series

INFocus provides exclusive insights and trends from experts and leaders across the Asia Pacific region, exploring the forces, opportunities, and challenges shaping its future.

Go to series

Yasuyuki (Yash) Sekioka

Hello everyone. I’m Yasuyuki Sekioka, partner at Oliver Wyman, Tokyo office. Today, we are joined by Tomochika Ishii, CFA, general manager of Equity Investment Department, Nippon Life, who will be discussing with us Nippon Life’s stewardship activities. We also hope to discuss Oliver Wyman’s support in those endeavors. With that, Mr. Ishii, thank you for joining us. 

Tomochika Ishii

Thanks for having me. I appreciate this opportunity. I am excited to discuss Nippon Life’s stewardship activities, which have been ongoing for over ten years.

Yash

Thank you, Mr. Ishii, I look forward to it. First, could you tell us what kinds of initiatives you have undertaken in stewardship activities over the past ten years at Nippon Life? Particularly amid environmental changes, I would like to hear how you responded to those changes.  

Tomochika 

Over the last 10 years, in pursuit of economic progress, corporate governance reforms have made a lot of progress. In 2014, we saw the establishment of Stewardship Code for Institutional Investors, and the Corporate Governance Code for companies in the following year. Acting in tandem, these two pillars have consistently advanced governance reform.

In this context, as a responsible institutional investor, we announced in 2014 that we would adopt the Stewardship Code and, while strengthening our framework, have worked to enhance both the quantity and quality of our dialogues. Specifically, in terms of quantity, we currently conduct dialogues with around 700 investee companies across Japan each year, and we believe we are among the most active in Japan. 

The themes covered in these dialogues have also gradually expanded. At first, most revolved around topics of governance and finance, but starting from 2018, we gradually expanded to themes such as the environmental and society as well. In terms of what we aim for, we want to be an institutional investor that is chosen by companies as a valuable dialogue partner. With that in mind, we engage in dialogue with many companies. We accumulate insights and use them to improve the quality of our dialogues. By doing so, we hope to keep evolving our activities going forward.  

Yash

Thank you very much. We now have a good understanding of what Nippon Life has done in stewardship over the past ten years. 

In this project with Oliver Wyman, we analyzed the dialogue meeting minutes from over ten years of ongoing activities in Nippon Life, to determine whether Nippon Life has truly conducted high-quality dialogues and whether they have been properly received by the investee companies, and further based on those dialogues, whether investee companies have actually undergone behavioral change that could be confirmed externally as observable outcomes. We supported you by analyzing and confirming these points.

To make our analysis as objective as possible, we used generative AI. First, we trained AI to understand Nippon Life’s investment philosophy. Then, we trained it to recognize what constitutes a higher-quality dialogue, and we had it read all the dialogue minutes to assess whether those dialogues were of high-quality and were creating impact. When launching our project, where there any particular expectations you had of Oliver Wyman? 

Tomochika 

We chose Oliver Wyman as our partner for three reasons. First, we have worked together with Oliver Wyman on stewardship projects over the past two years, which meant Oliver Wyman already had a deep understanding of our activities and philosophy. 

Secondly, even on a personal level, I really had the feeling that we were able to share the same goals and the level of quality we were aiming for — even before the concept stage of this project. 

Thirdly, the project was truly challenging for us, and we were expecting to hit many roadblocks along the way. Within that context, we expected Oliver Wyman would likely provide appropriate and useful solutions. Those were our expectations, those were the three reasons. 

In this project, you met those expectations. In fact, you exceeded them in two ways, and I would like to expand upon these. The first one, as you mentioned earlier, was the analysis done on the dialogues’ quality. This is where most time was spent. The nature of this endeavor on our end was to instruct generative AI on what constitutes an effective dialogue, and to get it to learn and assimilate accordingly across our 8,830 records so that it could appropriately assess the dialogues’ quality. That was the process. That is why the training and prompts were extremely important.

We started with a 300-record sample and iterated by trial and error. When you do trial and error with generative AI, while it excels at understanding context, it can also interpret our statements too favorably, so during the sample-checking stage, we performed tasks such as excluding such cases. In the end, the prompt’s contents came to roughly the equivalent of ten A4 pages. Also at that time, Oliver Wyman provided very practical solutions in our view. That was the first point.

Regarding the second point, we needed to assess, through generative AI, changes in corporate behavior. The procedure may look simple, but it is not so simple as we make various requests to the companies we invest in. For example, quantitative requests on the financial side or various qualitative requests such as enhanced disclosures. Given that context, the question was: how do we manage this systematically?

So first, we had generative AI to go through the requests from the dialogue minutes — over 10,000 of them— – and have it read all of them reliably, classify them into quantitative or qualitative requests, and pre-load and cross-reference with financial data on the quantitative side and disclosure materials such as securities reports and integrated reports on the qualitative side. By pre-loading these into the system environment, we could match our requests with the system’s data, and have the generative AI assess changes in disclosure documents and financial data. 

We asked your team at Oliver Wyman to perform those tasks, but to be honest, we were half-skeptical whether generative AI could really make such sound judgments. However, you managed to build the system environment admirably and made reliable judgements possible, which truly surprised us. Thank you very much.

Yash

Thank you very much. Through this project, looking back at Nippon Life’s stewardship activities over the past ten years, were there any new insights or discoveries? Could you share these? 

Tomochika 

In this analysis, the proportion of dialogues that were assessed as good in terms of quality and company response increased year by year, which we view positively. However, as expected, there were differences depending on the theme of the dialogues. For example, for dialogues on financial topics, relatively speaking, the proportion rated “High”, “Good”, or “Effective” was higher, whereas for themes like environmental and society, the proportion was relatively lower.

Each of our stewardship officer handles just under 100 investee companies per year. In that context, in order to raise the proportion of effective dialogues going forward, we felt we need to further consider the way in which we should explore our future activities. 

Regarding changes in corporate behavior, the proportion of cases where change occurred has increased year by year. Having said that, this is not to claim that the effect is solely due to our dialogues. Companies are influenced by many other factors, for example policy and requests from other institutional investors. Change can also arise from the company’s own corporate initiative. Among all these influencing factors, we see general alignment between our requests and the direction that companies take in terms of behavioral change. We see consistency in these results. We also found that when we conducted effective dialogues and received positive responses from companies, the proportion of change in corporate behavior was relatively higher. This confirmed that our activities were going in the right direction. 

Yash

Mr. Ishii, thank you for your valuable comments. Lastly, could you share Nippon Life’s view of stewardship, as well as your own vision, for future activities and initiatives? 

Tomochika

We believe that companies themselves must recognize the need for improvement and change of their own accord in order to improve corporate value in a sustainable manner. That is what we hope for. Recent discussions on Japan’s corporate governance reform focus on improving companies’ income sustainably and to that end, we will continue to engage with our investee companies. We believe many of our requests will take about 3 to 5 years, or longer, to achieve our requirements. With that in mind, we intend as a long-term-oriented institutional investor, to continue making requests to companies and support them in their initiatives. That is what we would like to do going forward. 

Yash

That was very clear. Mr. Ishii, thank you very much for today. 

Thank you for watching. For more information on this topic, please visit Oliver Wyman’s LinkedIn page and make sure to follow us for more updates.