← Cases
JapanConsumer goods2024 – ongoing

Sega Sammy Holdings

Image-generation AI and survey analysis at Sega Fave's toy business

Sega Fave's toy unit said its image-generation AI raised design proposals about 100-fold

WorkingCompany's own claim

Problem

The toy business needed more design ideas for product planning. It also had tens of thousands of customer survey responses, including free-text comments, to analyze.

Approach

Sega Sammy Holdings said on 7 May 2024 that Sega Fave's Toys Company had deployed a generative AI environment. An image-generation AI trained on its own product images produced design drafts close to existing products, and designers refined them. A second AI analyzed the survey responses. The environment runs on Platform Equinix.

Results

100倍increase in the number of design proposals, per Sega SammyCompany's own claim[1]
80%efficiency gain in survey analysis work in a proof of concept (about 80%), per Sega SammyCompany's own claim[1]

Takeaway

Training the image model on the company's own product images kept drafts close to existing products, and designers still did the final refinement.

Categories

Innovation type
Process innovation
AI technology
Generative AI
Evidence
Company's own claim

Sources

Source summaries

[1]Cloud Watch (Impress)21 Jun 2024

The article covers Equinix's announcement of the AI execution environment for the Sega Sammy Group. It reports the toy unit's image-generation AI, the 100-fold rise in design proposals and a proof-of-concept efficiency gain of about 80% in survey analysis. The results are Sega Sammy's own.

Full details are in the original sources.

Added 2 Nov 2026 · Updated 2 Nov 2026
Open case page ↗