Just thinking about ChatGPT or Midjourney won't keep up with the generative AI trends currently occurring in each industry.
In this book, we'll explore how leading companies are using generative AI to build their businesses and monetize them.
This book is a collection of business success stories utilizing generative AI, covering projects and services currently underway across various industries and companies.
We've collected successful cases of generative AI businesses across various industries, including chatbots in financial institutions, virtual experiences in the beauty industry, anomaly management in manufacturing, virtual design simulations, new drug discovery and clinical testing in the bio industry, and document creation and review in specialized fields such as medicine and law.

The industry has already created a lot of value with existing methods called “analytical AI.”
However, most analytical AIs are developed to perform specific tasks, making them difficult to apply to a wide range of situations or new situations.
On the other hand, ‘generative AI’ overcomes these limitations and can be applied to various situations.
So, while analytical AI is closer to a knowledgeable strategist with exceptional data analysis and decision-making skills, generative AI is closer to an artist who creates novel and creative content.
These characteristics are clearly evident in the fields of application.
Analytical AI is suitable for business intelligence, financial modeling, and predictive analytics, while generative AI is more widely used in creative fields such as art, design, and content creation.

--- p.11

Currently, most books focus solely on technical explanations, tool usage, and prompt utilization, and it is difficult to find books that cover business success stories that apply generative AI in real-world industrial settings.
As a result, the industry professionals I met were eager to increase their value by leveraging generative AI, and they were also curious about business applications.
When planning a new project, the first thing a company does is to look for so-called "advanced or successful cases."
However, there are not many widely known examples in the field of generative AI.
This is partly because generative AI is relatively new, but also because companies are reluctant to share their success stories publicly.

--- p.17

AI can handle repetitive, labor-intensive tasks like drafting job ads describing the job description, analyzing candidate responses, and producing reports comparing candidates' skills to the job openings.
Using ChatGPT, recruiters can resolve these issues much faster than before.

--- p.48

The world's biggest advertisers, from food giant Nestlé to consumer goods company Unilever, are already using generative AI to create and deliver ads.
WPP, the world's largest advertising agency in the UK, is partnering with consumer goods companies to use generative AI in their advertising campaigns.
You can create ads for just 5-10% of the cost of traditional advertising, and you no longer need to travel to the North Pole or Africa to film them.

--- p.66

LG Group plans to widely apply the use of an industrial autonomous AI agent called EXAONE across its production lines in electronics, chemicals, bio, and telecommunications.
It is possible to predict and detect equipment shutdowns and even detect anomalies that may occur during component assembly.
This can be said to be an example of a smart factory that has the triple benefit of increased productivity, reduced costs, and improved quality.

--- p.82

Morgan Stanley, an investment banking and asset management firm, launched a generative AI assistant based on GPT-4 for its wealth management advisors and support staff in September 2023.
This is the Morgan Stanley AI Assistant (AI @ Morgan Stanley Assistant).
This tool provides quick access to approximately 100,000 research reports and documents. The AI ​​assistant aims to provide quick and in-depth insights from massive piles of reports and documents for wealth management advisors.
It also reduces the time spent on various research tasks, allowing us to devote more time and focus on our customers.

--- p.118

Generative AI is driving innovation in healthcare, playing a role in a wide range of fields, from new drug development and clinical diagnosis to the synthesis and management of medical data.
Not only does it rapidly discover new drug candidates, it also creates new drug candidates by combining existing drugs in new ways.
Additionally, the lack of datasets required for new drug clinical trials is being filled with data created using AI.

--- p.126

For law firms that handle a lot of documents, such as case law research, litigation prediction, contract review and analysis, and document automation, legal tech is perceived as a time-saving service. This is because automation using AI's natural language processing and machine learning technologies can significantly increase efficiency.
Legal services utilizing legal tech are rapidly spreading overseas.
Already, large law firms and accounting firms are leveraging this technology to streamline labor-intensive tasks.
A number of legal tech startups have also emerged.
There are reports that a large domestic law firm has adopted AI lawyers, and a startup has also launched that uses AI to automatically prepare simple litigation documents.

--- p.142

Generative AI chatbots, including ChatGPT, are being integrated with automotive navigation systems, AI voice assistant systems, and driver assistance systems to enhance the driving experience and safety.
As generative design and AI are integrated into the automotive design process, automotive designs will evolve into more personalized and optimized forms.
By analyzing various driving environments and user preferences, we suggest the optimal body shape and interior space configuration accordingly.

--- p.172

Decorify is a virtual room styler provided by Wayfair, an American online retailer that primarily deals with furniture and interior products. It is a service that uses generative AI to help users design their homes and shop for furniture.
You can create various interior images with furniture arranged in a room, reflecting user feedback.

--- p.181

Pixar introduced generative AI to overcome the limitations of the existing animation production process.
As a result, the level of animation was raised while production time and cost were reduced by automatically generating storylines, background images, character designs, music, etc. for animated films.

--- p.196

Phobos has developed two AI tools called 'Adelaide' and 'Vertie'.
Launched in 2023, Adelaide uses generative AI to search and personalize Forbes articles relevant to readers' interests.
Prior to this, in 2019, we developed Bertie, an AI-based content management system that provides contributors (article writers) with features such as topic recommendations, draft article writing, and grammar checks.

--- p.244

L'Oréal has been the first to discover new trends in cosmetics and develop products that meet consumer needs.
L'Oréal has launched L'Oréal Beauty Genius, a virtual makeup try-on tool equipped with a generative AI chatbot using a large database collected from many countries around the world.
It uses personal photo information to check skin tone and condition and recommends appropriate cosmetics or makeup methods.

--- p.254

Similar problems arise in the field of natural language processing.
Some natural language processing systems tend to process occupational titles like "doctor" as strongly associated with masculine characters, and to process titles like "nurse" as strongly associated with feminine characters.
This is because the gender bias inherent in the training data has been learned.
In the field of image generation, bias is also shown, with cases where a male image is generated for the word 'CEO' and a female image is generated for the word 'kitchen'.

--- p.265

The CEO and CTO must work closely together. The CEO provides the CTO with the necessary resources and support and continuously monitors the project's progress.
The CTO regularly reports to the CEO on technical progress and potential risks. If data quality issues arise during the AI ​​model training process, the CTO explains the severity of the issue to the CEO and discusses the need for additional resource allocation or strategic adjustments to address the issue.
Successful AI application development is possible when the strategic vision and strong support of a CEO with insight and mid- to long-term foresight are combined with the technical leadership of an experienced and capable CTO.

--- p.279

Concerns about external transfer of customer data can be particularly serious in areas like the legal industry, which handles sensitive civil and criminal information about customers.
As a countermeasure, some argue that LLMs should continue to be used while enforcing very strict security rules and monitoring. However, companies skeptical of this approach are adopting sub-large language models (sLLMs) to alleviate security concerns while reaping the benefits of AI technology.

--- p.290

On-device AI works completely differently from LLM-based AI like ChatGPT.
Large-scale LLMs perform calculations based on computing resources such as servers in remote data centers via the cloud and then receive the results.
However, on-device AI processes data on the local device without sending it to the cloud.
That is, it collects and calculates information on its own without sending the information to the server.
It can be said that AI has been made lighter to be implemented within the device.

--- p.293