AI Shifts to the RAG Era, with 50% Engaged
– A Survey on the Use of Generative AI Conducted with 302 Companies and 402 Individuals, Showing a 20-Point Increase in Overall Adoption Compared to the Previous Survey

Exa Enterprise AI Inc (Minato-ku, Tokyo; Representative Director: Takuma Oue; hereinafter “Exa Enterprise AI“), a group company of ExaWizards Inc., conducted a survey at the Generative AI Seminar held May 27th and 29th, 2024. 302 companies and 402 individuals responded to the survey, with 50% of respondents indicating that they are currently engaged in internal data collaboration (RAG*), demonstrating a shift to the RAG era of generative AI.

*RAG (Retrieval Augmented Generation) allows generative AI to process and analyze customer-specific data to provide more accurate answers.

☑ Key Points
・50% of respondents are engaged in RAG. 40% show interest in RAG, indicating that it is a prerequisite.
・Utilization of individual generative AI continues to increase, with “Level 5 (daily use)” showing a steady increase. AI introduction to companies has expanded rapidly to nearly 60%.
・30% stated “internal data collaboration (RAG)” as a requirement to establish utilization. The higher the utilization rate of generative AI within the company, the greater the interest in RAG. 40% of respondents face challenges related to data management.

☑ Utilization of Generative AI in Individual Business (“Level 5 [daily use]”) Increased to 40%.
Utilization of generative AI in business has been classified into five levels: Level 1 (no interest), Level 2 (some interest), Level 3 (trial use), Level 4 (occasional use), and Level 5 (daily use) since the first survey in April 2023.
In the first survey, fewer than 10% of respondents answered “Level 5 (daily use).” When we conducted the same survey in May 2024, that percentage had increased to nearly 40%. This was a 4-point increase over the third survey (conducted December 2023) and was the largest increase among the different levels.

☑ 50% RAG Engagement
According to this survey, 50% of respondents are currently engaged in RAG, and about 40% show interest. Among them, 4% feel actual effects from using this tool. The increase in the percentage of Level 5 generative AI utilization indicates a correlation with RAG utilization within organizations, suggesting a shift to an era where RAG will be essential to generative AI.

☑ 40% of Respondents Face Data-Related RAG Challenges
The survey also revealed the challenges of RAG utilization. Many respondents mentioned date-related challenges, such as “required information not being digitized” and “data being in various formats.” Organizations with high generative AI utilization reported issues such as “insufficient pre-processing causing low accuracy” and “high expectations from users and management.”

☑ Generative AI Adoption is Rapidly Expanding
Generative AI adoption has expanded since the December 2023 survey, becoming nearly 60%. This widespread adoption has increased collaboration and sharing within organizations, leading to an increase in Level 5 utilization.

☑ In 70% of Companies, Less Than 30% of Employees Utilize Generative AI
Improving the utilization rate within organizations remains an important issue. This survey shows that the percentage of employees who “almost never use” generative AI has decreased, and the percentage of “30% of employees” at the company who use generative AI has increased. Companies with 100% utilization rates have increased the number of measures promoting utilization in their organization from 1.8 to 2.4 compared to the previous survey. Over 50% of companies are working on “prompt sharing,” indicating proactive efforts to promote utilization. Measures include “prompt sharing” and “sharing good internal utilization examples.” For details, please refer to the survey report.

☑ Internal Data Collaboration (RAG) Ranks First in Strategies to Establish Generative AI
To establish the use of generative AI, over 30% or respondents answered, “internal data collaboration (RAG),” making it the number one strategy. When companies responded that the utilization rate within the organization was “80% of employees” or higher, the percentage of those considering RAG utilization essential increased to 40%.

☑ Comments from Exa Enterprise AI: Representative Director: Takuma Oue
We have conducted the “Survey on the Use of Generative AI” four times since the public release of ChatGPT to understand the status of the generative AI market. One new finding from the latest survey is the prevalence of “RAG,” which allows collaboration on internal data. Our product, “exaBase Generative AI,” also has a RAG data collaboration feature that has received favorable feedback for providing accurate answers through generative AI. We also provide features to visualize productivity and support for internal establishment to improve utilization rates within organizations, allowing our clients to make their businesses more efficient.
The fact that over half of the companies surveyed have adopted generative AI is expected to enhance utilization rates. Executives who use generative AI accelerate its use within their organizations, which creates good utilization examples and allows internal departments to support one another.
Introducing RAG as a strategy for establishing generative AI is expected to be the key to improving business efficiency in the future.

☑ About exaBase Generative AI
Since its launch as a paid service in June 2023, exaBase Generative AI has been used by over 40,000 users.
It takes into account the security and compliance issues that are challenges for corporate use of generative AI, and provides features such as monitoring usage, registering prohibited words, and more for administrators. It also estimates the level of productivity improvement for each user, which can be visualized through graphs on the screen.
It is also possible to utilize generative AI based on the company’s own data. Users and administrators can upload their own files and use them for dialogue and generation.
For more information about exaBase Generative AI, please visit the following link:https://exawizards.com/exabase/gpt

【Exa Enterprise AI Profile】
Company Name: Exa Enterprise AI Inc.
Location: F, Sumitomo Realty & Development Mita Twin Building East, 4-2-8 Shibaura, Minato-ku, Tokyo
Established: October 2023
Representative Director: Takuma Oue
Business: Planning, development, and sales of products and services utilizing technologies such as generative AI to improve business operations and productivity
URL: https://exawizards.com/eai/

【ExaWizards Corporate Profile】
Company Name: ExaWizards Inc.
Location: F, Sumitomo Realty & Development Mita Twin Building East, 4-2-8 Shibaura, Minato-ku, Tokyo
Established: February 2016
Representative: Makoto Haruta, Representative Director & President
Business: Industrial Innovation and the Resolution of Social Issues via AI Service Development
URL: https://exawizards.com/en/

＜Media Contact＞
Exa Wizards Inc. Public Relations: publicrelations@exwzd.com