Kyushu Helps Small Companies Turn AI Experiments into Business Reform

In a meeting room attached to Hakata’s bus terminal, a group of Kyushu small-business owners will be asked to write an unusual declaration. Not how many hours artificial intelligence might save. Where those hours will go. Into new sales? Product improvement? Employee training? Faster service for customers? The Kyushu Bureau of Economy, Trade and Industry calls it a declaration of the “destination of time saved.”

The phrase cuts directly into a weakness of corporate AI programs. A summary can be produced faster, yet the recovered minutes disappear into another meeting or another manual transfer. Unless management chooses what the time is for, convenience does not necessarily become performance.

Next Leap AX—formally, the Management Transformation × AI Utilization Program—opened applications on July 28. It is intended for owners, executives and future executives at Kyushu SMEs. About 20 companies will be selected, generally with two people from each. Participation is free. Three sessions at Hakata Bus Terminal on September 15, October 5 and October 29 will move from diagnosing the business to defining a use case and then producing an execution plan. Applications close August 31.

Designed to go beyond “we learned about AI”

Regional digital support has long taken three familiar forms. A seminar provides knowledge. A subsidy helps buy equipment. A specialist offers advice. All three matter. But when the knowledge, tool and advice remain separate, adoption often ends as a one-off project.

Next Leap AX attempts to connect them as one management process. The first workshop separates what AI can do from the company’s actual problems. Managers assess their current position and declare the destination of any time saved. The second session uses cases and hands-on AI experience to turn the problem into a specific idea. The third defines the implementation theme, the process for spreading it inside the company, the people required, relevant personnel systems and the organization needed to receive outside expertise.

After the workshops, the program office will hold follow-up interviews and, for interested companies, present candidates with AX expertise. Those candidates may include members of the Fukuoka Prefecture CXO Bank. There is no introduction fee through the program, but a contract with a matched specialist will be paid. Public support carries a company through diagnosis and design; sustained implementation still requires the company to make its own investment decision.

Stage	Date	Management problem	Required output
Session 1	Sept. 15, 4–6 PM	Do not confuse AI’s capabilities with the company’s problem	Current position and a destination for saved time
Session 2	Oct. 5, 4–6 PM	Turn a useful demonstration into a business idea	A concrete AI use case
Session 3	Oct. 29, 4–6 PM	Design execution, talent, systems and internal expansion	Execution plan and talent requirements
Afterward	Follow-up interviews	Fill capabilities the company cannot hold internally	Optional connection to AX specialists

Kyushu’s industrial history is not a history of buying technology

The idea of adapting an outside technology to local work is not new in Kyushu. The government-owned Yawata Steel Works began operations in 1901. Imported German technology did not simply fit Japanese raw materials and production conditions. Furnaces, chimneys and coke-making processes were modified before mass production took hold. Industrialization came not from buying equipment alone, but from redesigning the process around local materials, skills and management.

After the war, Kyushu broadened its industrial base from steel and chemicals into automobiles and semiconductors. The semiconductor concentration known as Silicon Island has never consisted only of giant factories. Regional firms provide equipment, components, materials, maintenance and logistics beneath them. The renewed wave of semiconductor investment around Kumamoto has again made the region visible, but much of its implementation capacity still sits inside smaller suppliers.

AI differs from a blast furnace or semiconductor plant. It looks inexpensive and can be started by one employee. That accessibility is its advantage and its danger. Tools can multiply department by department without a companywide purpose, leaving subscriptions and risk to accumulate. This is why the Kyushu program places the business problem before the tool.

National data show the small-company gap

IPA’s “DX Trends 2026” found that 58.0% of Japanese companies had adopted or were testing AI. Implementation weakened as company size fell. Among firms with 100 or fewer employees, 34.6% were interested in AI but had no specific plans. The share pursuing DX at any level was 98.7% among companies with more than 1,000 employees and 41.6% among those with 100 or fewer. Companywide or departmental data use in the smallest group stood at 38.6%.

That gap cannot be dismissed as conservative local management. In a small company, the owner may simultaneously handle sales, hiring, cash, suppliers and customers. There may be no dedicated IT department, data owner or lawyer to check AI risk. Time for experimentation is scarce, and there is little capacity to absorb an investment that fails.

Smallness can also shorten the distance from decision to action. When the owner, the frontline and the customer are close, it may be easier to change one workflow from end to end. It matters that Next Leap AX targets management and expects roughly two participants per firm. One owner may fail to carry the field. One employee may lack authority to remove an approval. Two people can put authority and operating knowledge into the same conversation.

A small Fukuoka distributor found part of the answer early

SoNet is a Fukuoka company that sells factory-automation equipment. It had no dedicated IT department. More than a decade ago, it adopted a cloud-based sales system to reduce the burden of circulating Excel files for customer records and paperwork. To make the habit stick, its president even offered yakiniku dinners as a reward for employees who reliably entered their information.

The purpose is more revealing than the software. The 2005 Fukuoka offshore earthquake knocked an in-house server from a shelf and strengthened the company’s view of the cloud. Later, digitization allowed an employee who moved to Okinawa for family reasons to keep working remotely instead of leaving. In the company’s telling, the value of DX was not simply higher sales. It was becoming a company people did not have to quit.

That is an early example of finding a destination for time and flexibility. Efficiency was not used only to reduce headcount. It was redirected into retention, analysis, new-customer development and software the company could eventually sell. Small-business AI reform begins with the same choice.

Kitakyushu has been building a four-stage bridge

Next Leap AX did not appear from nowhere. Since 2020, the Kitakyushu DX Promotion Lab has organized its support around four stages: generate interest, prepare, practice and transform the business. Demonstrations and seminars create interest. Education for executives, managers and frontline leaders—plus a one-stop consultation service—prepares companies. Subsidies support practice. Awards and public cases spread successful transformation through the region.

Between 2020 and 2023, the one-stop service handled more than 300 consultations, while subsidies and related programs supported more than 150 digitalization or DX efforts. The sequence matters. A company can enter before it knows how to describe its problem and remain connected until an implementation becomes a business case.

One example recognized by the lab was Nishihara Shoji Holdings. It developed a system to improve administration and collection routes in its waste-management business, then offered the service to peers across Japan. A tool that solved an internal problem became an external product. The destination of efficiency was revenue and a change to the industry.

This suggests three levels of regional transformation. First, replace paper with digital information. Second, make the process faster. Third, turn the resulting data and capability into a new service. Next Leap AX is a test of whether small firms can carry AI into that third level.

An outside specialist is not a magician

The matching element is one of the program’s most realistic features. A small company is unlikely to hire full-time experts in AI, data, workflow design and security. A project-based, side-job or fractional specialist can provide the required knowledge for a defined period.

But an outside specialist cannot bring clarity into a company that has not defined its problem. “We want to do something with AI” tends to produce a tour of tools. Which customer is waiting for which outcome? Which decision needs which data? Who stops the system when it is wrong? Until management can answer those questions, it cannot define the person it needs.

That is why the third workshop includes talent requirements, personnel systems and a receiving structure. A strong specialist still fails if the company will not provide data access, protect an employee’s project time or name an executive who can decide.

Five ways to measure reform

Measure	Weak version	Business-reform version
Time	Minutes saved on an AI task	Days removed from the whole lead time
Revenue	Number of texts or images generated	Quotation speed, win rate and customer value
Profit	Whether the subscription is cheap	Margin after rework, scrap, inventory and outsourcing
People	Number trained or prompts written	Overtime, retention, learning time and decision quality
Customers	Whether a chatbot was installed	Wait time, first-contact resolution and satisfaction

If these baselines are not recorded before implementation, a polished plan on October 29 will not produce an answer the following spring. AI generates outputs. The company must supply the starting point.

Kyushu is testing management commitment, not AI

Twenty companies are not enough to transform Kyushu. Three workshops totaling six hours will not overturn a corporate culture. A contract with outside talent costs money. Each company will still face data cleaning, changes in authority and employee anxiety.

Yet the program focuses on something worth watching. It refuses to count learning sessions and trials as the final result. It connects AI to saved time, revenue, profit, organization, personnel systems and the kind of outside help management is prepared to receive.

The history of Yawata Steel Works offers a useful warning. Technology does not become an industry when it arrives. It becomes an industry when it is adjusted to the work, when processes change, people learn and surrounding firms connect. Generative AI is no exception.

Before the August 31 deadline, an owner considering the program should not begin by choosing an AI product. The first question is which work the company should no longer continue. The second is who should benefit from the time recovered. Without those answers, an experiment remains an experiment. With them, six hours in Hakata could become the point at which a small company changes direction.

Sources and reporting notes

1. Kyushu Bureau of Economy, Trade and Industry, Next Leap AX participant recruitment, July 28, 2026.
2. SME Support Japan J-Net21, Next Leap AX event listing, for dates, venue and cost.
3. IPA, “DX Trends 2026” and the main report, for AI, DX and data use by company size.
4. IPA Local DX Promotion Lab, Kitakyushu’s integrated SME support model, including consultation totals and the Nishihara Shoji case.
5. IPA DX SQUARE, SoNet’s employee-centered DX.
6. IPA, Manabi DX Quest 2025 recruitment, for the development of practical, company-linked learning.
7. City of Kitakyushu, history of the Imperial Yawata Steel Works.
8. Cabinet Office, Japan’s Second Artificial Intelligence Basic Plan, provisional translation, as of July 14, 2026.

Editor’s note: The program is recruiting and has not begun; this report does not predict or guarantee outcomes. “Silicon Island” is historical context for Kyushu’s industrial concentration and does not mean participation is restricted to semiconductor companies. The publisher supplied the exchange rate; its UTC timestamp was converted to Japan Standard Time.