MHongyan Meng
Japanese English Chinese

Business Design / Priority 1

Initial design for an AI search and analysis report business

The first target is not a large system contract. It is a small analysis report that can be updated every day. Public examples build trust and connect to consultation, contract analysis, and AI-search implementation.

Free public outputBuild trustDaily memos, charts, short analysis
Small revenueMonthly reportsMarket, technical cases, agricultural data
Contract workAI search buildPDF, CSV, specs, research material

Value provided

Specialized documents, news, price time series, technical cases, agricultural data, and old source code can be searched, organized, and converted into short analysis reports. The strength here is the combination of reading old materials/source code and implementing AI, search, and data visualization.

Public pages do not expose confidential information. They show themes, methods, and decision material only. That makes them sales evidence: visitors can understand that document organization and AI search can be requested.

Public project themes

SDD/LSI-style search technology

Experience in agricultural document retrieval, concept search, and matrix decomposition can be connected to modern RAG, paper search, and internal knowledge search.

Agriculture and field data

AgrInfo, XML/XSD, sensors, barcodes, and CSV management can be converted into agriculture-DX and field-measurement reports.

Image recognition and federated learning

YOLO, Docker, Flower, and experimental-environment setup can become verification procedures, experiment logs, and pre-introduction diagnosis.

Legacy C/Java repair

Old C, Java, Windows GUI, SWT, and Tomcat assets can be read and diagnosed for repair risk, migration strategy, and reusability.

30-day execution plan

  1. Week 1Stabilize the AI Analysis Lab market and crypto data refresh, and leave one short observation memo every day.
  2. Week 2Create public sample articles in three themes: technical cases, financial risk, and agricultural data.
  3. Week 3Narrow inquiry paths to document organization, AI search, and analysis reports, then publish price guides.
  4. Week 4Put the URL on GitHub, X, freelance profiles, business cards, and proposals, and aim for one small consultation.

First sales menu ideas

Existing PDF/CSV organization

Classify material and prepare search keywords, summaries, and CSV conversion direction.

From 5,000 JPY

Pre-AI-search diagnosis

Diagnose which documents can be turned into RAG material, which are difficult, and what data should be prepared.

From 30,000 JPY

Monthly analysis report

Organize markets, technical cases, agriculture, or measurement data in the same monthly format.

From 10,000 JPY / month

Important cautions

  • Financial data must not be presented as trading recommendations. Keep it as observation, verification, and risk-management records.
  • For old or client materials, summarize only what can be public and avoid proper nouns or confidential details.
  • Do not aim for a large development contract at the beginning. Build evidence through small organization, research, and report work.
  • Even a small 100-yen-per-day experiment becomes service quality if the reason and result are recorded.

Conclusion

The first priority is the AI search and analysis report site. It can be operated by one person, becomes public evidence, and every small daily improvement becomes sales material. From there it can expand to crypto-risk detection, agriculture/weather forecasting, and legacy AI/C repair.