How Generative AI is Transforming the Strategic Value of “Non-Digital Assets”

The Moment the Definition of “Assets” Changes

The proliferation of generative AI is rewriting the very definition of corporate “assets.” Information that was previously difficult to recognize as an asset—non-digital and unstructured—is beginning to demonstrate its true value for the first time, thanks to AI.

The direction indicated by the recent news of Fukushima Minpo’s entry into the generative AI business and the efficiency improvements in debt collection operations by Ippu Senkin and View Card share a crucial commonality. It is the “AI-driven re-evaluation and utilization of existing non-digital assets.” The archived articles of a local newspaper and the know-how of debt collection were previously “sleeping assets” whose digitalization costs did not justify their value. However, the advent of generative AI is dramatically changing this cost structure.

The Challenge of a Local Newspaper: Article Archives as the Foundation for a “Regional AI”

Behind Fukushima Minpo’s entry into the generative AI business lies the existence of its vast accumulated archive of local article data. The articles of a local newspaper constitute a unique database that records the region’s history, industry, culture, and challenges in detail. However, until now, they were merely stored as searchable PDFs or in databases.

The situation changes with the emergence of generative AI, particularly models that can be trained and fine-tuned with proprietary data (utilizing RAG: Retrieval-Augmented Generation technology). An AI trained on this article data can provide diverse services, such as creating marketing materials for local businesses, supporting municipal policy planning, and multilingual translation of tourist information.

The crucial point is that this asset is “impossible for others to imitate.” Nationwide AI services may be strong on general knowledge, but they cannot understand the deep context or historical background of a specific region. The “deep local knowledge” possessed by local newspapers can become a powerful competitive advantage in the generative AI era.

Practical Steps: Uncovering Your Company’s “Sleeping Assets”

What managers and back-office leaders should do first is take an inventory of their company’s “non-digital assets.” Specifically, the following items can be used as a checklist.

• Documents & Records: Past sales reports, customer service records, internal manuals, meeting minutes
• Human Knowledge: Tacit knowledge of experienced employees, customer service “style,” troubleshooting rules of thumb
• Information Attached to Physical Assets: Equipment maintenance records, negotiation history with suppliers, quality inspection data
• Communication Records: History of customer inquiries, complaint handling records, back issues of company newsletters

The first step to making these assets usable for AI is not “digitalization” but “evaluating the potential for structuring.” There is no need to digitize everything immediately. It is important to prioritize which assets could contribute to operational efficiency or the creation of new services.

The Efficiency of Debt Collection Operations Reveals the Value of “Process Assets”

The case of Ippu Senkin and View Card sheds light on another type of “sleeping asset.” That is the “know-how of the business process itself.” Debt collection operations involve tacit knowledge accumulated over years, including legal knowledge, timing of negotiations, and nuanced approaches tailored to each customer.

This initiative, utilizing the generative AI “Court,” achieves standardization and efficiency by training the AI with this tacit knowledge. Specifically, the AI learns from past successful and unsuccessful debt collection cases, optimal approaches based on customer attributes, and supports their application to new cases.

The key lesson from this case is that the success of AI implementation depends not on “mere tool introduction” but on the “formalization of business know-how.” Many companies fail in AI adoption because they underestimate this formalization process.

Cost Considerations and Specific Implementation Approaches

The costs and approaches for utilizing a company’s non-digital assets with AI can be organized as follows.

Small-Scale Start (Approx. $200 – $330 USD per month): Begin by using business plans like ChatGPT Enterprise or Claude Team, uploading company data to answer questions. At this stage, no large-scale data formatting is done; existing PDF or Excel files are used as-is.

Medium-Scale Customization (Approx. $660 – $2,000 USD per month + initial development costs): Building models fine-tuned with proprietary data or introducing RAG systems. This stage requires data preprocessing (anonymizing personal information, unifying formats, etc.). Often involves utilizing external AI development partners.

Large-Scale In-House Development (Initial investment of $6,600 USD or more): Developing proprietary AI models or deep integration with existing business systems. This stage requires securing engineering talent in-house or a long-term contract with a trusted development partner.

For many small and medium-sized enterprises, a realistic approach is to start with a “small-scale start” and gradually increase investment while confirming ROI. Based on my own experience, I recommend starting with a project that trains AI on past successful case studies from the sales department and uses it as a training tool for new sales staff.

The Essential Value of “AI Advisory” Services

The true value of the “Free Feasibility Diagnosis” offered by the ‘Hands-on AI Implementation Support (AI Advisory)’ service introduced here lies precisely in this “discovery of your company’s sleeping assets.” External experts objectively assess the value of assets your company may not have noticed and their potential for AI integration.

However, as a manager, it’s important to note that the “diagnosis” provided by these services remains merely a presentation of possibilities. Actual implementation success depends on subsequent internal data preparation and business process review. Rather than blindly trusting the diagnosis results, it’s crucial to formulate a realistic plan by comparing them with your company’s resources (time, personnel, budget).

The Key to Success: Data “Quality” and “Context”

The biggest challenge when utilizing non-digital assets with AI is ensuring data “quality” and “context.” Past documents often have inconsistent formats, terminology that changes over time, unstated implicit assumptions, and many other elements understandable to humans but difficult for AI.

The solution is a step-by-step approach like the following:

1. Selecting Sample Data: Start with the most valuable and relatively well-organized data.
2. Adding Context: Add metadata (when, by whom, for what purpose it was created, etc.) to the data.
3. Creating a Glossary: Create a glossary defining terms and abbreviations unique to the company and train the AI on it.
4. Gradual Learning: Begin learning with a small amount of data, and gradually increase the training data while verifying the output results.

This process takes time, but once the data assets are organized, they demonstrate value not only for AI utilization but also in diverse scenarios such as standardizing internal knowledge, training new employees, and ensuring business continuity.

“Asset Re-evaluation” as a Management Strategy

The advent of generative AI redefines the source of corporate competitive advantage. Elements previously unrecognized as assets—such as “deep local knowledge,” “tacit process know-how,” and “years of customer service expertise”—are now becoming visible and productizable for the first time through AI.

The next move as a manager is to systematically uncover and evaluate your company’s “sleeping assets.” The criteria for judgment at that time are the following three points:

• Scarcity: Is it an asset that competitors cannot easily imitate?
• Ease of AI Integration: To what extent can it be utilized with current technology?
• Business Impact: Can it become a new revenue source, not just improve efficiency?

The Fukushima Minpo case, while appearing to be about a specific industry like regional media, actually contains a lesson applicable to all companies. It is that “the unique data and know-how accumulated over many years can become your greatest weapon in the generative AI era.”

When considering AI implementation, before first looking for the latest external tools, take the time to look deeply within your own company. There, you may find the seeds for your next growth, unique to your company alone.