CyberAgent, Inc. has announced the launch of “CA-GAS” (CyberAgent-Generated Ads Structure), a new advertising framework designed to optimize ad delivery based on the diverse search intents of modern users.
Addressing the Complexity of Modern Search
As user search behavior becomes increasingly fragmented and complex, there is a growing need for systems that can efficiently manage and operate a high volume of creative assets tailored to individual intent.
Moving away from traditional keyword-centric delivery models, CA-GAS leverages the strengths of Generative AI in “intent comprehension” and “content generation.” The system designs ad groups starting from the creative itself, enabling granular delivery that aligns with the specific context of each user.
Leveraging Proprietary AI Technology
The core of this new structure integrates two of CyberAgent’s established AI systems:
- Kiwami Yosoku LP: A system that predicts the effectiveness of landing pages.
- Kiwami Yosoku TD: A tool that automatically generates optimized ad descriptions (Title/Description).
By combining these technologies, CA-GAS achieves highly precise matching between the user’s query and the promotional content.
Proven Performance and Future Outlook
In preliminary testing, the service demonstrated significant results, with the number of conversions (CV) increasing by 1.34x while maintaining the existing Cost Per Acquisition (CPA).
While the service is currently available to companies producing dedicated landing pages, CyberAgent plans to expand the offering to businesses utilizing their own corporate websites. The company remains committed to leveraging advanced AI technology to maximize advertising performance for its clients.
