Hakky List of Products and Services

This document explains the AI-driven knowledge circulation solution "Hakky Handbook for Business," which aims to prevent the personalization of sales activities. To break the current situation where the quality of deal preparation depends on the abilities of individual representatives, it introduces the importance of establishing an infrastructure that centrally manages information in a format that is easy for both AI and humans to handle. By linking AI with a knowledge base and operating two loops of different natures, it is expected to elevate the overall level of the organization. [Contents] ■ Local Loop: A cycle that improves the closing rate and speed of individual projects ■ Wide Loop: A cycle that enhances the overall strength of the organization and achieves standardization *For more details, please download the PDF or feel free to contact us.

• others

This document explains that the introduction of AI serves as a pressure device that promotes the verbalization of organizational culture. It provides essential know-how for successfully articulating the organization during AI implementation. There are many instances where the entire team struggles to articulate why they are doing something during the AI introduction process. This is why it is crucial to verbalize the criteria for judgment, verification, reproduction, and explanation. This document discusses three disruptions faced in AI-based business design, such as business stagnation due to insufficient verbalization. [Contents] ■ Business stagnation due to insufficient verbalization of judgment criteria ■ Issues of accountability arising from unclear responsibility boundaries ■ Weakening of communication due to the breakdown of mutual adjustment culture *For more details, please download the PDF or feel free to contact us.

• others

This document introduces a transformation roadmap consisting of three steps: "points, lines, and surfaces," for welcoming AI agents into organizations as members rather than tools. The transition to an AI agent organization is realistically advanced through "points, lines, and surfaces," and to mitigate the risk of runaway behavior, an approach of "discarding the assumption of inherent goodness and binding through systems" is adopted. While accumulating insights in the knowledge base 'Hakky Handbook for Business,' horizontal deployment will be promoted. Know-how for quantitatively evaluating and improving quality, speed, cost, and risk will be obtained. [Contents (partial)] ■ Transformation roadmap to an AI agent organization ■ Accumulation and sharing in the knowledge base 'Hakky Handbook for Business' ■ Measures against runaway risks ■ KPI setting across four axes *For more details, please download the PDF or feel free to contact us.

• others

This document explains "AI Ready Knowledge Management," which organizes and manages the knowledge that an organization possesses so that AI can autonomously access, accurately interpret, and utilize it. By reading this document, you will gain know-how to activate accumulated knowledge assets through AI, creating value such as preventing knowledge loss during employee turnover, improving training speed, and fostering innovation. In recent years, with the rapid development of generative AI and AI agents, there is a growing demand to optimize business operations by leveraging proprietary data. However, traditional explicit knowledge accumulated with the assumption that humans will read it is not necessarily in a form that is easy for AI to use. Therefore, efforts to prepare the knowledge in a state that AI can appropriately interpret and reuse have become essential. [Contents] - Solutions to issues such as data silos and insufficient rule design - Three foundations: "Meaning Structure," "Data Structure," and "Knowledge Utilization Infrastructure" - Support for building an AI Ready environment *For more details, please download the PDF or feel free to contact us.

• others

This document explains the cost discipline and withdrawal design to resolve the gap between investment and outcomes in AI investments. Currently, many AI initiatives are abandoned before they go live, making the gap between investment and outcomes visible. What is questioned during this settlement period is not the pursuit of new models, but the judgment of which to continue and which to stop. This document clarifies the causes of the gap between investment and outcomes and presents three designs to overcome this issue. 【Contents】 ■ Causes of the gap between investment and outcomes (going live without designing a cost structure, experiments without a definition of success) ■ Three designs to overcome the gap (establishing GO/NO-GO criteria based on numbers, cost control through structure, focusing on a few use cases) ■ Support provided by our company (creating GO/NO-GO criteria, optimizing cost structures, delineating between withdrawal and continuation, etc.) *For more details, please download the PDF or feel free to contact us.

• others

This document is an explanatory material summarizing the "five decisions" necessary for successfully implementing AI. It explains the core know-how needed to prevent project stagnation and transition to the execution phase. It discusses the background in which "designing decisions" is required rather than technical skills, including the establishment of standards for companies that have achieved an 80% commercialization rate for PoC, decision-making structures, and the redesign of business processes. [Contents (partial)] ■ "Designing Decisions" that determine the success or failure of AI implementation ■ Definition of success criteria and withdrawal conditions before starting ■ Changes to the organizational operation model based on AI assumptions *For more details, please download the PDF or feel free to contact us.

• others

We would like to introduce a case study of developing a foundation and dashboard for utilizing various data that has been dormant within Progrit Inc. Due to the lack of an integrated data foundation, the data aggregation necessary for documents was being done manually. Additionally, as this was their first time building a full-scale foundation, they were looking for a partner to assist with the overall design. As a result, we built the infrastructure and pipeline for the data foundation, as well as a dashboard for visualization, in a short period of time. We designed a user-friendly dashboard that does not require high literacy levels and also supported the selection of tools to reduce operational burdens. 【Case Overview】 ■Challenges - Data aggregation necessary for various documents was being done manually each time. - Cross-business data analysis was taking up a lot of man-hours. - The accumulated data was not being utilized sufficiently. ■Effects - Regular visualization of manual data aggregation through BI, reducing aggregation man-hours. - Established a governance foundation considering audits and permissions in three and a half months. *For more details, please download the PDF or feel free to contact us.

• Other Software

We would like to introduce a case study of Blue Mo Securities Co., Ltd. regarding the transition to a data infrastructure that balances security and data utilization. As the company’s services grew, the volume of data processing increased, revealing issues with the existing data infrastructure. It became difficult to consider dependencies, making it challenging to understand the impact of one change on other processes. In response, we implemented automatic masking of personal information, achieving immediate anonymization. We built an environment where data could be utilized safely while simplifying permission management, significantly reducing management workload while maintaining security standards in the financial industry. [Case Overview] ■ Challenges - The complexity of data processing dependencies made it time-consuming to respond to changes. - It was difficult to balance strict personal information management with data utilization and operations. - Ensuring safety and usability in collaboration with partner companies. *For more details, please download the PDF or feel free to contact us.

• Other Software

We would like to introduce a case study in which we built a customer support chatbot system utilizing a large-scale language model for O.M. Network Co., Ltd. The company has been deploying a shift management system, but due to the centralized approach at headquarters and reliance on a rule-based bot, they were unable to provide adequate individual support. By trialing advanced AI models and leveraging accumulated conversation history, we achieved high-precision user support. Additionally, we established a system within the company to implement product improvements. [Case Overview] ■Challenges - Inadequate individual support for actual end users - Desire to build a flexible chatbot to improve customer satisfaction *For more details, please download the PDF or feel free to contact us.

• Other Software

We would like to introduce a case study of developing a data analysis chatbot for television production at Data Stadium Inc. The company was considering a system that would allow for easier analysis and was interested in value provision that combines a data infrastructure with a large language model. Therefore, we developed a tool that can acquire data related to baseball. This tool can display the desired information in a suitable format in text form, eliminating the need for complex analysis instructions. We focused on usability, including features that address specific needs such as correcting inconsistencies in notation. 【Case Overview】 ■ Challenges - Considering a system for easier analysis - Desired development of a tool to acquire baseball data in chatbot format ■ Effects - Improved usability through prompt engineering *For more details, please download the PDF or feel free to contact us.

• Other Software

We would like to introduce a case study of a Proof of Concept (PoC) for diagnostic support AI that assists with diagnostic discrepancies in a beauty clinic for Virginia Corporation. The company was considering incorporating diagnostic support AI into their existing system, but they were unable to achieve practical accuracy through in-house development. Therefore, they defined requirements based on the actual counseling flow and successfully built an AI that meets the needs of the field. To properly train the AI on diagnoses that can lead to differing judgments, they conducted multiple interviews to redefine annotation standards that could be understood even by non-medical professionals. As a result, they succeeded in improving the accuracy of the AI model without increasing the amount of data. [Case Overview] ■ Challenges - They want to incorporate diagnostic support AI into their cloud business support system in the future. - They want to support the diagnostic discrepancies arising from the experience of physicians. - They could not achieve practical accuracy through in-house development. *For more details, please download the PDF or feel free to contact us.

• Other Software

We would like to introduce a case study of a dialogue-based manga recommendation service developed using generative AI for Shueisha Inc. The company was considering a service to recommend over 5,000 manga titles they own, tailored to users' preferences. We provided an early prototype for the challenging requirement of defining a conversational experience. This allowed for a smooth transition to production without significant setbacks. We incorporated essential character information and the desired conversational experience into the prompts, enabling unique responses. [Case Overview] ■ Challenges - Considering a service to recommend manga based on user preferences. ■ Results - Reached a total of 200,000 recommendations within two months of launch. - Achieved unique response expressions through character-specific prompt engineering. *For more details, please download the PDF or feel free to contact us.

• Chatbots

We would like to introduce a case where we developed and provided an evaluation system for "contest application attractiveness" using generative AI for the Public Recruitment Guide Co., Ltd. The company was manually evaluating approximately 2,000 to 3,000 new public recruitment applications each month with a few staff members, which was a significant burden in terms of labor. Additionally, there were issues with inconsistencies in evaluation criteria depending on the reviewer, leading to fluctuations in evaluations. In response, we verbalized the subjective evaluation criteria through interviews and achieved a highly accurate public recruitment evaluation using AI. This allowed for the complete automation of evaluations for approximately 3,000 applications per month, resulting in evaluations that are more accurate and consistent than those made by humans, significantly contributing to operational efficiency. [Case Overview (Partial)] ■Challenges - To convey the enjoyment of public recruitment to users and revitalize the contest industry. - To enhance the added value on the site through automated evaluations by AI. - To manage the burden of evaluating approximately 2,000 to 3,000 new public recruitment applications each month. *For more details, please download the PDF or feel free to contact us.

• Other Software

We would like to introduce a case where TAPP Co., Ltd. developed and provided "Akari AI," a conversational AI that supports the anxieties of investment beginners. The company wanted to incorporate AI into its sales process in line with the trends of the times and achieve a web customer service unique to the AI era. However, they were unsure about the specific implementation methods and how to integrate it into their actual sales flow, and they had concerns about the market perspective and approach to their first AI development. In response, our company developed a dedicated conversational AI that learned from actual consultation cases accumulated during the company's asset management seminars and the question trends of users. This AI enables customer interactions equivalent to those of in-house sales representatives, regardless of time or place, significantly contributing to the realization of new web customer service in the AI era. [Case Overview] ■ Challenges - Methods for realizing "web customer service in the AI era" - Specifics on how to integrate into the "sales process" - Concerns about "cost and feasibility" regarding first-time AI development ■ Effects - Successfully increased the conversion rate to about three times that of immediately after release, reaching the expected KPI. *For more details, please download the PDF or feel free to contact us.

• Chatbots

This document is a collection of solutions summarizing the services provided by Hakky Co., Ltd., including AI product development and generative AI solutions. It includes detailed information on various AI-related solutions such as "annotation," "generative AI: automatic article generation," and "AI agent-equipped product development." We also introduce approaches for quickly carrying out everything from requirements definition to implementation and operation, so please take a moment to read through it. 【Contents (partial)】 ■ Agile AI Development ■ Requirements Definition ■ Annotation ■ Generative AI: Automatic Article Generation ■ Generative AI: Chatbots ■ Generative AI: Automatic Document Generation *For more details, please download the PDF or feel free to contact us.

• Other Software

Hakky Inc. is a company specializing in data and AI, engaged in various businesses to update everything related to "products" using the power of data and AI. We support the development of the necessary infrastructure for data analysis to utilize our own data, as well as the operation of tools and security. To enhance the value of products through the use of machine learning, we provide total support from requirement definition to PoC and production. We also offer automated analysis tools such as 'aigleApp' and qualitative research tools like 'FindVox'. 【Business Contents】 ■ Support for Data Infrastructure Development ■ Support for AI Product Development ■ Support for Data Utilization *For more details, please download the PDF or feel free to contact us.

• Other Software