Introduction

In recent years, generative AI has evolved remarkably, and its potential goes beyond simply improving business efficiency; it has the potential to have a huge impact on transforming the very business models of companies. In 2023, generative AI has been reported almost daily in Japan, and it is no longer a passing fad, but is now recognized as a full-fledged technology trend that will affect corporate competitiveness.

Strategic implementation of generative AI can not only revolutionize existing business processes, but also lead to the development of new products and services, improved customer experiences, and even the creation of new markets.

In this article, we will introduce some success stories, focusing on the points to keep in mind when introducing generative AI and creating a system to make it take root throughout the organization. We hope that this article will be helpful in drawing up a strategy for how to turn generative AI into an engine for your company's growth.

Current status and challenges of using generative AI in companies

The use of generative AI has made great strides overseas, particularly in the United States and China, and many companies are accelerating their adoption as part of efforts to strengthen their competitiveness. Meanwhile, in Japan, only 45% of companies are using generative AI in their operations, meaning there is still a lot of room for growth.

One of the factors causing this difference in adoption rate is the policy for introducing generative AI. Many Japanese companies generally build a general-purpose ChatGPT for internal use, and each business department explores how to use it. In contrast, overseas companies, including those in the United States, place emphasis on an approach that directly incorporates generative AI into specific business processes, and aim to establish it at a practical level by transforming the entire operation. It is believed that these differences in implementation style have a significant impact on the actual utilization rate and results.

In fact, after interviewing multiple business divisions, the following two issues emerged:

The first challenge is that the chatbot-type UI of generative AI is a general-purpose UI, and it is not clear how it can be applied to which tasks, making it difficult to understand how it should actually be used. When considering implementation across an organization, there are many cases where a lack of specific usage scenarios leads to trial and error, and the benefits of implementation cannot be maximized.

The second challenge is that it is difficult to design the "prompts" to be input to the generative AI, and if the appropriate instructions are not given, the expected results will not be obtained. For example, if a marketing team wants to use generative AI for seminar planning or content generation, they need to carefully design what instructions (prompts) should be given to get the best output. Trial and error of these prompts is surprisingly time-consuming, and many departments found this to be a hurdle in the early stages of implementation.

Initiatives to solve problems (co-creation with users)

To resolve these issues and effectively utilize generative AI in business operations, it will become increasingly important for the department promoting the introduction of AI and the business department to work together to explore ways to utilize the technology and build optimal usage scenarios through "co-creation" efforts.

In this article, we will introduce two specific measures that we are taking.

1. Developing optimal input methods in collaboration with users

At our company, the department promoting the introduction of AI and the business department are considering the use of generative AI through the following four processes: "Identifying and prioritizing business operations," "Confirming input and output," "Developing input screens and setting prompts," and "Deploying to business operations."

First, in "Confirming Input and Output," you organize the information items currently used in your work and the results you ultimately want to achieve. Next, in "Developing Input Screens and Setting Prompts," you define and develop input items that are in line with your work based on the information you organized, and set prompts taking into account the desired output.

By going through this process, business department staff will be able to obtain the desired output using generative AI by simply entering the necessary information into pre-designed input fields, without having to interact with a chatbot.

For example, in our company's marketing department, the sales department inputs the title and summary of the seminar, the target persona, etc. into Excel, and then considers the seminar summary to be posted on the web based on that information. As a result of introducing the above-mentioned system into this business process, the generation AI was able to generate multiple proposals for seminar summaries for the web based on the content entered into Excel, making it possible to significantly reduce the workload of the person in charge.

2. Accumulation and use of prompt output results

By operating the system developed through the above process, the results output by the generative AI are evaluated, and specific examples of outputs that are suitable for business and those that are not are gradually accumulated. By systematically accumulating knowledge about such ideal outputs and establishing a mechanism to reflect this in the output of the generative AI, it becomes possible to produce more accurate results that are in line with the needs of the company and its operations.

Summary

There are many challenges to the introduction and adoption of generative AI, but the key to overcoming these is a "co-creation" initiative in which the department promoting the introduction of AI and the business department work together. It is important to repeatedly consider and develop optimal input methods and output results so that business departments can make the most of generative AI, and to further evolve the AI based on the results.

Our goal is to fully realize the potential of generative AI through this co-creation and revolutionize business processes. We hope to continue sharing specific implementation methods and use cases to contribute to improving your business operations.

If you are considering using generative AI, please feel free to contact us. Let's create the business processes of the future together.