The “AI Co-Creation” Initiative by Local Businesses and Government Points to the Next Generation of Business Strategy

Simultaneous “AI Co-Creation” Pilots in Local and Central Government

The Toyokawa Chamber of Commerce in Aichi Prefecture and the central government’s Ministry of Agriculture, Forestry and Fisheries. At first glance, these two entities seem unrelated, yet they both began moving towards “generative AI utilization” and “co-creation with AI agents” almost simultaneously. This is not a mere coincidence. It is a significant sign that the use of generative AI is evolving from a “personal productivity tool” into a “new foundation for inter-organizational collaboration.”

The Toyokawa Chamber of Commerce held seminars to accelerate generative AI adoption and new business creation for local companies. Meanwhile, the Ministry of Agriculture conducted an “AI x Policy Planning Hackathon,” demonstrating next-generation administrative practices where AI agents and humans “co-create.” What these cases share is an attempt to reposition AI from a tool to “use” into a partner to “collaborate with,” integrating it into the very processes of organizational decision-making and value creation.

Why SMEs and Startups Now Have the “Advantage”

Recent news has pointed out that startups and SMEs have an “advantage” over large corporations when it comes to generative AI. Based on my practical experience, this assessment is spot on. There are three reasons for this.

Speed and Flexibility of Decision-Making

In large corporations, introducing AI tools often requires internal proposals and security reviews, sometimes taking months. In contrast, SMEs and startups can implement tools on the same day based on a single decision by management. When I introduced Claude and ChatGPT into my operations, I weighed the cost (approximately $140 USD per month) against the expected benefits and made the decision that very day. This speed difference becomes a decisive advantage in the current climate of accelerating AI technology evolution.

Fewer Constraints from Existing Systems and Processes

Large corporations have core systems and strict business processes built up over many years. Integrating AI poses a major hurdle in connecting with these “legacy” systems. Conversely, SMEs can design AI-first workflows almost from a blank slate. For example, it’s possible to perform “zero-based design,” automating everything from sales reports to customer management and invoice creation through a series of AI agents.

Lower Cost to Build an “AI Agent System”

My team has built a system of 32 AI agents with specific roles, delegating tasks from social media posts to contract reviews and code generation. Building such an “AI agent system” can be achieved at a fraction of the cost and time compared to large-scale system implementations in big corporations. All that’s needed are subscriptions like ChatGPT Plus or Claude Pro (approximately $15-$20 USD per ID per month) and basic automation skills to connect them.

How to Start “AI Co-Creation”: 3 Practical Steps

Some business leaders might see the Toyokawa and Ministry of Agriculture cases as “distant stories.” However, co-creation with AI agents can start as soon as tomorrow. Here are three concrete steps.

Step 1: Accumulate Your Company’s “Tacit Knowledge” in an AI-Readable Format

The premise of co-creation is that the AI understands your business. First, digitize and organize the following materials for storage.

• Reports on past successes/failures
• Frequently asked customer questions and their answers
• The “commitments” or “philosophy” behind product/service development
• Industry-specific glossaries and transaction practices

Store this information in cloud storage like Google Drive or Notion. The key is to have it in a searchable text format, not as PDFs or images. This allows the AI to learn your company’s “context” and provide more accurate proposals and analysis.

Step 2: Appoint “Dedicated AI Agents” Specialized for Specific Tasks

Instead of expecting a “do-everything AI,” create specialized AI agents for each business function. For example:

• Market Analysis Agent: Regularly monitors competitor websites, social media, and news to create reports.
• Drafting Agent: Creates drafts for sales emails, proposals, and blog articles, matching your company’s tone and style.
• Operational Efficiency Diagnostic Agent: Analyzes documents describing internal workflows and proposes points for automation.

Give each agent a clear role (“You are an expert in…”) and access to the internal materials accumulated in Step 1. Using features like ChatGPT’s “Custom GPT” or Claude’s “Projects,” these can be built for free or for a few thousand yen per month.

Step 3: Introduce Regular “AI Co-Creation Meetings”

The most important step is to establish a regular forum for dialogue between humans and AI. Even 30 minutes once a week is sufficient. Run the meeting with the following agenda:

1. Reports from AI: Each agent presents analysis results or proposals for their assigned tasks.
2. Human Feedback: Correct aspects the AI alone cannot judge, such as “This analysis doesn’t consider on-the-ground realities” or “The cost assumption in this proposal isn’t realistic.”
3. Next Instructions / Learning: Based on the feedback, provide new instructions or add learning data for the AI.

Repeating this process transforms AI from a mere tool into a “strategic advisor” that deeply understands your business. The essence of the “co-creation” aimed for in the Ministry of Agriculture’s hackathon lies precisely in this continuous dialogue and mutual learning.

Pitfalls to Watch For and Cost Considerations

It would be disingenuous to offer only optimistic projections. Here are three points business leaders should be particularly mindful of when advancing AI co-creation.

Managing Information Security and Intellectual Property Risks

Inputting confidential information directly into public AI is risky. There are two countermeasures. First, for highly sensitive processing, use OpenAI’s API called from your own servers with settings that prevent data storage. Second, if you must use public AI, use “anonymized data” where specific figures and proper nouns are replaced with dummy values. When I automated contract reviews with AI, I also used a two-step process: analyzing with placeholder names and amounts, followed by a human finalizing the document.

Initial Investment and Expected ROI

Building a full-fledged AI co-creation system requires some initial investment.

• Tool Costs: Subscriptions for ChatGPT Plus, Claude Pro, code-assistance AI (like Cursor), totaling approximately $70-$140 USD per month.
• Labor Cost / Time Investment: Internal resource time (from the leader or an assigned person) for system setup. About 5-10 hours per week for 2-3 months.
• Training Cost: Basic training on AI utilization (online courses costing a few tens of thousands of yen).

On the other hand, the expected benefits go beyond just time savings from automating routine tasks (my company saved 1,550 hours annually). Data analysis by AI can uncover previously unnoticed customer insights or operational bottlenecks, potentially providing hints for new business ventures. This is precisely why the Toyokawa seminar focused on “new business creation.”

Balancing “AI Reliance” and “Human Judgment”

The greatest concern is uncritically accepting AI proposals and neglecting final human judgment. AI makes proposals based on past data patterns, but it is humans who create the future. Especially for ethical judgments and strategic decisions based on a company’s long-term vision, humans must take responsibility. AI is the “best strategic advisor,” but the “commander” must always be the human leader.

The Future Belongs Not to “Organizations That Use AI” but to “Organizations That Co-Create with AI”

The support for regional businesses in Toyokawa and the policy planning hackathon by the Ministry of Agriculture are ahead of the curve in the paradigm shift of AI utilization. Future competitive advantage will depend not on “how much AI tools are used,” but on “how well AI can be nurtured as a strategic partner for co-creation.”

This change presents an excellent opportunity precisely for SMEs and startups, which have faster decision-making and fewer constraints from existing systems. The steps to start today are clear: systematize internal knowledge, “hire” task-specialized AI agents, and nurture them through regular dialogue. Beyond this lies a future where local businesses and SMEs, despite constraints on human resources, can generate new ventures with speed and creativity that rivals even large corporations.

AI is no longer just a tool for the IT department. It is an essential resource for all leaders, redefining business strategy itself.