AI Transparency Policy

1. Basic Policy

We regard AI not as a replacement for human judgment, but as an operating base that extends human judgment and practical experience. On that basis, we hold to the following principles when using generative AI in our work.

(1)Transparency: We explain what we use AI for and what people decide, to the extent reasonably possible in delivering our services.

(2)Security: We prioritize safeguards for client personal information and confidential information.

(3)Quality accountability: We carry out human review and quality control of AI-assisted deliverables. Any legal liability relating to a deliverable provided under a client engagement is governed by the applicable agreement.

(4)Continuous review: We review our AI practices and governance as the technology evolves.

(5)Practicing what we advise: What we recommend to clients is grounded primarily in what we have tested in our own operations.

2. AI Tools and Related Services We Use

The principal AI tools and related services we use are listed below. As a general rule, we select business plans, APIs, or equivalent configurations that allow administrative control, and we run them with settings under which submitted data is not used to train or improve a model.

(1)Tools in regular use

Large language models:

• Claude (Anthropic, PBC, United States)
• ChatGPT (OpenAI, L.L.C., United States)
• Gemini (Google, L.L.C., United States)

Operations and CRM:

• Google Workspace (Google LLC, United States)
• HubSpot (HubSpot, Inc., United States)
• Notion (Notion Labs, Inc., United States)

Website and design:

• Webflow (Webflow, Inc., United States)
• Figma (Figma, Inc., United States)
• Canva (Canva Pty Ltd, Australia)

Automation:

• Google Apps Script (Google LLC, United States)
• n8n (n8n GmbH, Germany)

(2)Tools we may use depending on the engagement

Research and analysis:

• Perplexity (Perplexity AI, Inc., United States)
• Make (Celonis SE, Germany)

Development and coding support:

• Cursor (Anysphere, Inc., United States)
• Claude Code (Anthropic, PBC, United States)
• GitHub Copilot (GitHub, Inc., United States)

Image, video, and voice generation:

• Midjourney (Midjourney, Inc., United States)
• Runway (Runway AI, Inc., United States)
• ElevenLabs (ElevenLabs Inc., United States)

(3)Tools under consideration

3. Work Supported by AI and Decisions Reserved for People

We draw a clear line between work that may be supported by AI and decisions that remain with people.

(1)Work that may be supported by AI, subject to human review

• Preparing first drafts (content, reports, proposals, and similar material)
• Summarizing, organizing, and classifying data
• Translation, proofreading, and style consistency
• Research and information gathering
• Generating images and graphic assets
• Generating code and automation scripts
• Automating repetitive tasks

(2)Decisions always made by people

• Strategic recommendations and advice on management decisions
• Conclusions from financial analysis and the basis for investment decisions
• Final review of deliverables containing client or confidential information
• Execution of contracts and other material decisions
• Material communications with clients
• Fact-checking and quality review of AI-assisted output
• Governance and compliance judgments
• Setting exit criteria and assessing business risk

AI improves operating efficiency, but professional judgment and quality control are carried out by our people. This division is a basic principle of how we manage quality.

4. Handling of Client and Confidential Information

When using AI tools, we handle client personal information and confidential information in accordance with the following rules, as a general practice, except where a client has directed or consented otherwise under an individual agreement.

(1)We do not enter client personal information or confidential information into an AI tool in identifiable form.

(2)Where information is entered into an AI tool, we apply the following measures as needed beforehand, so that individuals and organizations cannot be identified.

・De-identification (replacing names, company names, and other identifying details)
・Summarization (generalizing specific figures and identifying details)
・Masking (concealing highly confidential elements)
・Sanitization (removing personal, financial, and similar information)

(3)Where we use an AI tool in our work, we select, as a general rule, a business plan, an API, or an equivalent configuration allowing administrative control, under which submitted data is not used to train or improve a model.

(4)Information obtained under a non-disclosure agreement with a client is handled with particular care in relation to entry into an AI tool.

(5)Because tool specifications, terms, and settings may change, we check the applicable conditions of each service at the time of use and manage that use appropriately.

5. Quality Control of AI-Assisted Output

We apply the following quality-control process to deliverables produced with the help of AI tools, appropriate to their nature and intended use.

(1)Fact-checking: We check the information, figures, and citations produced by AI against reliable sources.

(2)Context review: We confirm that what AI has produced fits the client's circumstances, industry, and problem.

(3)Quality review: A person reviews the logic, the professional standard, and the appropriateness of the text.

(4)Editing and correction: A person edits, adds to, or corrects the material as necessary.

(5)Final approval: The person responsible gives final approval before the deliverable goes to the client.

Deliverables are provided to clients only after this process. We do not relax our quality standards because AI has been used. Any legal liability relating to a deliverable provided under a client engagement is governed by the applicable agreement.

6. Copyright and Intellectual Property

(1)AI-assisted output and rights of use: We handle output generated using AI tools in accordance with the terms of each AI service. Before providing a deliverable to a client, we take reasonable steps to confirm that we hold the rights necessary for its intended use.

(2)Protection of client intellectual property: We exercise appropriate care so that entering client-provided information, materials, or intellectual property into an AI tool does not infringe the rights of third parties.

(3)Third-party rights: For output generated by AI, we carry out a reasonable review, appropriate to its nature and intended use, of possible infringement of third-party copyright, trademarks, and rights of publicity. Where a concern arises, we investigate the facts and take appropriate action.

7. AI Governance Framework

We refer to the following domestic and international guidelines and risk-management frameworks in working towards the safe use of AI tools, sound governance, and continuous risk management.

(1)Japan: the AI Guidelines for Business published by the Ministry of Economy, Trade and Industry and the Ministry of Internal Affairs and Communications / the Act on the Protection of Personal Information and the guidance of the Personal Information Protection Commission

(2)International: the OWASP Top 10 for LLM Applications / the NIST AI Risk Management Framework / the NIST Generative AI Profile

These materials inform our governance and risk management. Reference to them does not constitute a certification or a representation of compliance with every requirement of each framework.

8. Incident Response Relating to AI Use

Where a material incident or concern arises in connection with our use of AI tools, we act without undue delay, appropriate to the nature of the matter. Our response may include the following.

(1)Where a concern about disclosure of information arises: investigating the facts, notifying the individuals concerned and reporting to the relevant authorities where necessary, and considering and implementing measures to prevent recurrence

(2)Where a quality problem arises in AI-assisted output: identifying the problem and analyzing the cause, reporting to the client promptly and responding, and correcting or reissuing the deliverable

(3)Where an AI provider changes a service or suffers an outage: considering and implementing alternatives, assessing the effect on clients, and notifying them where necessary

For the detailed response flow, please also see Section 16 (Response to a Data Breach) of our Privacy Policy.

9. Supporting Clients in Their Use of AI

We make available to clients the AI methods we have tested in our own operations.

(1)Package 1-1: AI in-house readiness and P&L review — we put outsourcing spend, AI headroom, and governance risk into numbers.

(2)Package 1-2: AI security and governance review — we design the governance framework for using AI safely.

(3)Packages 2 to 4: strategy design, build, and automation — we build a marketing operation designed around the use of AI.

(4)Package 5: Ongoing improvement — as an external CMO or CAIO, we support monthly improvement in how AI is used.

What we recommend to clients is grounded primarily in what we have tested ourselves. This website is one example of that.

10. Changes to This Policy

We may amend this Policy in response to developments in AI technology, changes in law, changes in the tools we use, or changes in the wider environment.

An amended Policy takes effect when posted on our website. Where a change is material, we will give notice on the website by an appropriate method.

Where a tool we use is added or changed, we update the list in Section 2.

11. Relationship to Other Policies

This Policy supplements our Privacy Policy, Cookie Policy, and Terms of Service.

(1)Handling of personal information: please also see Section 8 (Handling in Connection with AI Tools) of our Privacy Policy.

(2)Cookies and tracking technologies: please see our Cookie Policy.

(3)Corporate information and trade secrets: please see Section 17 (Corporate Information and Trade Secrets) of our Privacy Policy, together with the non-disclosure agreement or services agreement entered into separately.

12. Contact

For questions concerning this AI Transparency Policy, please contact us:

For enquiries about AI advisory or assessment work, please use the inquiry form or the free consultation booking on our website.

Japan Consulting Inc.

Additional Rules

This AI Transparency Policy takes effect on May 20, 2026.
This AI Transparency Policy was amended on July 31, 2026, and the amended Policy takes effect on the same date.

Established: May 20, 2026

Last updated: July 31, 2026