The State of Generative AI Programming in 2025 — How No-Code is Transforming "Development Culture" and "Implementation Structure" —

Introduction: The era of no-code is evolving into "AI-driven programming"</h>

In the 2022 blog "The Era Where Anyone Can Be a Creator with No-Code AI", the focus was on "no-code technology" that allows creating apps and AI without specialized programming knowledge.

At that time, against the backdrop of a shortage of IT personnel, no-code was expected to be the "starting point for in-house development," and there were forecasts that the domestic market would reach a scale of 100 billion yen in fiscal 2023.
Three years have passed since then.
As of 2025, no-code is evolving by combining with generative AI, shifting from the traditional visually-driven construction to an AI-powered conversational programming approach. AI assists in design and writes code. And it is becoming commonplace for even non-engineers to be able to develop—. This structural change is significantly transforming companies' implementation capabilities.

▼Reference Blog
2023 AI Trend Forecast: The Era Where Everyone Becomes a Creator with No-Code AI.

1. Maturation of the No-Code Market and the Shift to "AI Programming"</h>

The no-code/low-code (LCNC) market continues to grow steadily. In 2022, there was a forecast that the market's CAGR (compound annual growth rate from 2020 to 2025) would be 24.4%.

Subsequently, according to the latest ITR survey, although the tone has been somewhat lowered with a forecast that the market's CAGR (2023 to 2028) will be 12.3% and that the market size in 2028 will be 1.8 times that of 2023, there is no doubt that the market itself continues to grow. No-code initially spread as a tool for startups and idea validation, but it is now evolving into a position as business application and development infrastructure within enterprises.

According to the August 2025 report by Index.dev, 41% of companies are already actively promoting in-house app development by non-developers (such as business departments), and those employees are developing an average of 13 business apps. This movement is changing the conventional notion that "programming = a specialized profession" and making the arrival of an era where "anyone can create apps through programming" a reality.

Furthermore, as of 2025, many no-code tools have integrated generative AI, so when users input requirements in natural language, the AI automatically proposes and implements workflows and code.
In other words, we are now in an era where "programming with AI by your side to assist you has become the norm."

These changes are transforming the very nature of development through "AI-assisted programming," redefining companies' implementation capabilities.

2. Structural Changes Brought by Generative AI: What Is the Standard Form of No-Code AI Programming?</h>

Around 2023, prompt engineering was still something "skilled people wrote." It largely depended on individual skills, and in reality, prompts were often shared and used within teams.

However, as of 2025, that structure has changed significantly. In no-code tools equipped with generative AI, users only need to convey their requirements in natural language, and the AI automatically generates optimal prompts behind the scenes and builds workflows and UIs automatically.
The concept of "writing prompts = programming" has merged within the GUI.

Additionally, the issue of individual dependency, which was a challenge with traditional prompts, has also been resolved. Excellent prompts are now shared and improved within internal libraries and knowledge bases, evolving into a stage where they are managed as "organizational knowledge assets." Furthermore, technologies such as Anthropic and LangChain have standardized mechanisms where AI analyzes its own outputs and optimizes prompts. In other words, AI tunes AI, while humans focus exclusively on design and verification—this is the standard form of "no-code AI programming" in 2025.

3. Challenges in the No-Code × Generative AI Era: How to Manage AI Programming</h>

Programming with generative AI has indeed dramatically increased development speed. However, behind this lies the unavoidable issue of "the safety of AI-generated code." Malfunctions caused by hallucinations, leakage of confidential information, copyright and licensing issues, and security vulnerabilities.
What many companies are currently facing is the practical challenge of "how to manage code written by AI."

● Limitations of Customization and Scalability
No-code tools are basically designed to combine pre-made parts, so there are limits to handling complex data processing and large-scale loads. Therefore, a hybrid type of AI programming that partially combines traditional code development is required.

● Dependency on Tools
While no-code tools are convenient, relying too heavily on a single platform can lead to a loss of flexibility in the future. To avoid being affected by specification changes or pricing model revisions, it is essential to make ease of data export and flexibility in integration with other systems key criteria when selecting tools.

● Security and Quality Assurance
Code and workflows generated by generative AI may contain oversights and vulnerabilities. Therefore, it is important to utilize AI-based code review and security inspection tools to establish a systematic checking framework. Not only development speed but also how well the verification and quality assurance processes are managed organizationally will determine the outcomes.

● The Flip Side of "Anyone Can Create"
As the base of development skills expands, what differentiates the quality of outcomes is design capability and understanding of the business processes targeted by the program. In generative AI programming, humans are required to have the ability to judge "what and why to design in a certain way." Companies have moved past the stage of increasing the number of "people who can create" and are now transitioning to the phase of nurturing "people who can design."

4. Five Practical Steps Companies Should Take</h>

As of 2025, companies that are achieving results by leveraging generative AI share a common trait: they have institutionalized AI utilization as a system and are enhancing its maturity.
According to VentionTeams' 2025 report, companies with high AI maturity have an ROI three times greater than those still in the experimental phase. Additionally, a BCG survey reports that 74% of companies that have introduced AI have not yet achieved value creation, indicating that what determines success is the design of the system and the institutionalization of its operation.

The five practical steps shown below are common points observed in companies that have established such a "mechanism to enhance the maturity of AI programming utilization."

1. Strategic Selection of Generative AI-Compatible Tools

The key to AI utilization lies in "choosing based on the intended operation." Evaluation criteria include flexibility of AI integration, security, future portability, and scalability. A design that easily incorporates operational rules such as data management and permission settings will determine the outcomes.

2. Expansion from Small-Scale PoC to Institutionalization</h>

If the evaluation criteria are biased only toward "speed" or "cost," projects tend to stop at the PoC stage. It is important to document and share the effects obtained from verification as reproducible processes and embed them into the mechanisms of the entire team.

3. On-site-led Development and Cultural Establishment</h>

On-site-led development by non-engineers is expanding, and a structure that connects the development and operations traditionally handled by engineers is required. By embedding the promotion and management of AI utilization as a culture throughout the entire company or organization, an environment is created where "people who can design correctly" can be nurtured.

4. Clarification of AI Usage Rules and Responsibility Scope</h>

Code and documents generated by generative AI tend to have ambiguous responsibility, so it is necessary to clearly define copyrights, modification responsibilities, and verification processes, and establish a system that treats AI outputs as managed assets. Incorporating AI usage logs into the audit process can minimize risks.

5. Multilayered KPIs</h>

It is important to shift the evaluation focus from measuring AI implementation results by "speed" and "cost reduction" to multilayered indicators that include quality, reusability, and risk reduction.

5. The Next 10 Years: Programming with Generative AI Becomes a "Competition of Implementation Capability"</h>

Generative AI combined with no-code is no longer just a "technology for creating programs," but has become a system for steadily building business processes and services and continuously improving them. The focus has shifted from "anyone can create" to "who can continuously and correctly operate it."

In an era where AI generates code, the question is not about tool selection but about the development culture and management capabilities themselves. While technology is shared, the systems for how AI is operated and managed differ from company to company. This "ability to design organizational structures" will be the key axis determining competitiveness over the next decade.

Furthermore, companies that fail to solidify their structure may face a future in 2035 where they "lose in competition despite using AI."

6. Conclusion: Generative AI Programming Evolves from On-Site Technology to a Management Foundation</h>

Generative AI programming has evolved beyond being merely an "on-site development method" to becoming part of the framework that supports management. It is no longer just about introducing tools, but about operating them as a system, embedding them as a culture, and designing them as a structure. This capability has become the axis that determines organizational competitiveness in this new era.

The era where "anyone can create" has now become a reality, not just an ideal. The next question is who will promote and manage this process, and how they will transform the results into business value.

Here, the key to success lies in whether the organization can design the cycle—where the field defines the issues, AI generates the programs, and the organization verifies and optimizes them—as an institutional system. This design capability depends on organizational culture and management skills.

Generative AI programming has evolved beyond being merely a "means of DX" and has become a foundation that supports the structure of enterprises. True transformation will begin only when organizations make "structural reproducibility" a key initiative.

7. Human Science Training Data Creation, LLM RAG Data Structuring Outsourcing Service</h>

Over 48 million pieces of training data created

At Human Science, we are involved in AI model development projects across various industries, starting with natural language processing and extending to medical support, automotive, IT, manufacturing, and construction, just to name a few. Through direct business with many companies, including GAFAM, we have provided over 48 million pieces of high-quality training data. No matter the industry, our team of 150 annotators is prepared to accommodate various types of annotation, data labeling, and data structuring, from small-scale projects to big long-term projects.

Resource management without crowdsourcing</h>

At Human Science, we do not use crowdsourcing. Instead, projects are handled by personnel who are contracted with us directly. Based on a solid understanding of each member's practical experience and their evaluations from previous projects, we form teams that can deliver maximum performance.

Generative AI LLM Dataset Creation and Structuring, Also Supporting "Manual Creation and Maintenance Optimized for AI"</h>

In addition to labeling for data organization and creating training data for identification-based AI, we also support structuring document data for generative AI and LLM RAG construction. Since our founding, manual production has been our main business and service, and we now also assist with organizing business knowledge and manual creation aimed at future generative AI and RAG implementation and utilization. We provide optimal solutions leveraging our unique expertise in the structure of various documents.

Secure room available on-site</h>

Within our Shinjuku office at Human Science, we have secure rooms that meet ISMS standards. Therefore, we can guarantee security, even for projects that include highly confidential data. We consider the preservation of confidentiality to be extremely important for all projects. When working remotely as well, our information security management system has received high praise from clients, because not only do we implement hardware measures, we continuously provide security training to our personnel.

In-house Support</h>

We provide staffing services for annotation-experienced personnel and project managers tailored to your tasks and situation. It is also possible to organize a team stationed at your site. Additionally, we support the training of your operators and project managers, assist in selecting tools suited to your circumstances, and help build optimal processes such as automation and work methods to improve quality and productivity. We are here to support your challenges related to annotation and data labeling.