Adobe GenStudio - Content Intelligence
Providing and explaining marketing features from content
Adobe GenStudio is a Generative AI application that allows marketing teams to create high performance content. I initially worked on a GenStudio Insights proof-of-concept to recommend content for Adobe Express . An A/B test showed +100% sign up CTA clicks, +65% marquee CTA clicks, +22% template clicks, +37% downloads, +33% MAU in express with my proof-of-concept recommendations. GenStudio Insights works by using AI to extract features studied in marketing and psychology research from advertisements and uses this predict what resonates with different audiences.
Going further, I worked on developing models to identify emotional drivers in marketing content through research across psychology and marketing. I have also led and mentored intern projects to improve the quality of models and make them explainable for generating controlled variations of ad copies. With the widespread adoption of LLMs, I built a RAG-based system to generate actionable marketing insights using explanations from the identified emotional drivers and performance attribution models.
Finally, I worked on LLM-based experience fragmentation for webpages or multi-pod ads to help identify self-containing pods that may have their own user interactions.