AI-Native MarTech – How we Built Feedbooster with Google's Gemini to Supercharge e-Commerce скачать в хорошем качестве

AI-Native MarTech – How we Built Feedbooster with Google's Gemini to Supercharge e-Commerce

Speakers: • Nafsika Bourma, Senior Digital Business Strategy Manager, Relevance • Giannis Stamatelos, Senior Cloud Engineer, Rocket Path #MarTech #Ecommerce #ProductOptimization #ContentAutomation #GoogleGemini #DigitalMarketing Nafsika and Giannis introduce FeedBooster, an internal tool developed to solve the "messy data" crisis in e-commerce. They argue that traditional methods of optimizing product feeds—manual entry or simple "if-this-then-that" rules—are unscalable and produce robotic content that hurts conversion rates. Key Highlights: ✅ The Scalability Bottleneck: Managing product content for thousands of daily-updating SKUs across different channels (SEO, Ads, various marketplaces) is impossible with manual resources or simple automation. ✅ Failure of Legacy Automation: Previous "rule-based" automation couldn't produce natural, native-language descriptions, leading to robotic content and lower conversion rates that undermined the purpose of automation. ✅ Enterprise AI Features: Dynamic Field Injection allows injecting product variables directly into prompts for context-aware descriptions. Prompt Library provides shared repository for pre-tested prompts ensuring team consistency. Pre-Flight Visibility shows estimated cost and operations before bulk jobs. ✅ Model Arbitrage for Cost Control: Sophisticated ability to "select the right tool for the right job"—deploy smaller, cheaper models for simple tasks (fixing titles) and reserve larger, expensive models for complex creative writing, optimizing P&L. ✅ Beta Results: Reached 95% feed quality. Ad disapprovals declined and campaign performance improved, directly linking data hygiene to revenue—proving "Gen AI isn't just theory, it's how we are winning auctions." 💬 Why Watch? Essential for e-commerce leaders, digital marketing professionals, and martech platform builders. This session demonstrates how AI solves real operational bottlenecks in product content management at scale. Learn how to balance AI cost with quality through intelligent model selection. 📌 Notable Insights: The "Prompt-First" workflow empowers non-technical e-commerce specialists to act as engineers, creating custom data fields using only natural language. Architecture on Google Cloud handles parallel processes for millions of products securely, demonstrating production-scale engineering beyond prototypes. Model arbitrage—using different sized models for different complexity tasks—is essential for managing the unit economics of AI at enterprise scale. Feed quality directly impacts ad performance and revenue—data hygiene through AI isn't just operational efficiency but revenue generation. About the GenAI Summit: Over 5,000 GenAI builders, technology executives, and political and business leaders convened on November 24 at the Stavros Niarchos Foundation Cultural Center for the 4th GenAI Summit, organized by 100mentors. Under the theme "From Chatbots to Agents," the Summit captured the evolution of GenAI: from simple Q&A systems to autonomous platforms that plan and execute complex tasks. European product "premieres" and participation from leading international companies cemented the GenAI Summit's position as Southeast Europe's most influential artificial intelligence conference. Founded in 2023, GenAI Summit has established itself as Southeast Europe's most influential AI event. It convenes researchers, entrepreneurs, and government leaders shaping GenAI's future, connecting scientific research with practical implementation across enterprises and organizations. For more information: genaisummitseeurope.com #MarTech #Ecommerce #ProductOptimization #ContentAutomation #GoogleGemini #DigitalMarketing #GenAISummit #Innovation #Technology #ArtificialIntelligence