Automating Semi-Annual Sales Promotion Data Aggregation of 30,000 Items and 50 Patterns with Generative AI -- A Sales Promotion/Campaign Management Company Embarks on "Skill-Based Aggregation" and Company-Wide Cost Optimization with Claude Code

Comix Inc. (Headquarters: Shibuya-ku Maruyama-cho, Tokyo; Representative Director: Akihiro Suzuki) announces that it has begun an initiative to automate data aggregation and documentation tasks, which occur semi-annually and involve approximately 30,000 items and 50 aggregation patterns, for its client companies engaged in sales promotion and campaign management, using generative AI (Claude Code).

The approach utilizes a method that allows tool construction through Japanese conversation alone, without programming knowledge, and proceeds in a phased manner: "Initial Execution → Skill Creation → Scheduled Execution (Full Automation)."

Background and Challenges

In the field of sales promotion and campaign management, it is essential to aggregate campaign results for each client and incorporate them into proposals and reports. However, in the marketing and advertising industries, the volume of data handled is enormous due to the complexity of digital advertising and the increasing demand for personalization. It has become virtually impossible for humans alone to optimize everything, making the automation of data processing, analysis, and decision-making by AI indispensable.

The client company in this case was also facing the following challenges:

- Aggregating a campaign database of approximately 30,000 items once every half-year.

- A wide variety of approximately 50 aggregation patterns.

- The task itself is standardized but requires significant manpower due to the large volume.

"The patterns are fixed, but manual work takes a long time" – this is precisely the area suitable for automation with generative AI, and there was an urgent need to improve the efficiency of standardized tasks that were reaching their limits with individual effort alone.

Provided Services

Support for "Skill-Based Aggregation" using Claude Code

Comix first prepared the client company's development environment (setting up Git/Node.js/VS Code, etc.) and conducted a demonstration of automatically generating and adding functions to spreadsheets with Google Apps Script using only Japanese instructions. After allowing them to experience that tools can be created through conversation alone, even without prior programming experience, they are now proceeding to the next stage of automating campaign data aggregation.

- Initial Execution via Conversation: Provide raw data and have the AI execute the entire process from aggregation to documentation, then fine-tune the output and create a skill.

- Skill Creation: Have Claude Code memorize successful processes as "skills," enabling the reproduction of the same quality output by simply providing data from the next time.

- Scheduled Execution (Full Automation): Aim for full automation in the future, where "processing automatically occurs when data is placed in a designated folder, and notifications are sent via Slack."

Inquiries regarding generative AI utilization support can be made here.

Inquiries regarding this article can be made here.

Features and Strengths

- Site-led Design: Business automation tools can be built through Japanese conversation alone, without programming knowledge.

- Organizational Asset Creation of On-site Know-how: Successful processes are memorized by AI as "skills" and can be reproduced by anyone.

- Phased Automation Approach: A reasonable design that starts with a small-scale implementation and aims for full automation in the end.

- Dual Promotion with Cost Optimization: Support that goes beyond operational efficiency to include a review of company-wide fixed tool costs.

Target Users and Usage Scenarios

Target Users

- Management / Department Heads

- Site Managers in Marketing and Sales Promotion Departments

- Information Systems / Administration Departments

Example Usage Scenarios

- Indirect Operations: Automation of standardized aggregation of vast amounts of data, report creation, and documentation tasks.

- On-site Operations: Analysis of campaign results and organization of data for planning future measures.

- Organizational Strength Enhancement: Organizing the strengths of each generative AI by comparing them to "personalities with roles" and establishing them within the company (e.g., "Claude the Commander" for structuring and system design, "ChatGPT the Secretary" for text refinement and consultation, "Gemini" for images and videos, etc.).

Future Outlook

Generative AI demonstrates its true value not by one-off efficiency improvements, but by "memorizing as skills and using them repeatedly." Standardized tasks are precisely the area where AI excels, and we accompany our clients not just by providing tools, but by enabling them to build automation themselves and delve into optimizing fixed costs.

First, we aim to make this aggregation task a successful internal case study and support the expansion of AI utilization throughout the company. We strive to turn efficiency improvements into "management impact" by supporting not only implementation but also operational design and human resource development.

Inquiries regarding generative AI utilization support can be made here.

Inquiries regarding this article can be made here.

--- Comix Inc. Company Profile ---

Headquarters Address: Kondo Building 2F, 15-4 Maruyama-cho, Shibuya-ku, Tokyo

Representative: Representative Director Akihiro Suzuki

Established: September 2007

Business Activities: Generative AI utilization productivity improvement support, sales support for BtoB businesses, referral of freelance professional talent, etc.

Web: https://www.comix.co.jp/

Comix Academy: https://www.comix.co.jp/academy/

【Contact】

TEL: 03-5459-5394 MAIL: n.adachi@comix.co.jp Contact Person: Comix Inc. Naruse Adachi