New Way Now: PUMA fields 75% of in-store customer questions with AI built on Google Cloud and NVIDIA
지금 보고 있는 영상
New Way Now: PUMA fields 75% of in-store customer questions with AI built on Google Cloud and NVIDIA
조회수 476
영상 설명 · 눌러서 펼치기
*Featured in this video:* Ivan Dashkov, Head of Emerging Marketing Technology, PUMA
*Executive summary:* PUMA teamed up with Google Cloud, NVIDIA, and LiveX to deploy Dylan, an innovative, AI-powered virtual assistant that delivers expert product advice to shoppers. Built on Google Kubernetes Engine and Cloud Run, the system dynamically scales to handle peak retail traffic. By taking an agile multi-model approach, PUMA uses Gemma for quick inquiries and routes detailed catalog queries to NVIDIA Nemotron. Dylan securely ingests PUMA's extensive product catalog, supporting over 100 languages to eliminate shopping barriers for international guests. The virtual assistant now autonomously resolves 75% of in-store customer inquiries, allowing store associates to prioritize complex interactions and providing shoppers with a seamless, conversational journey from product discovery to checkout time.
*Challenge:* As one of the world's premier sports brands, PUMA prides itself on delivering high-end experiences to both everyday consumers and globally-ranked athletes. Even common products like running shoes have become sophisticated tools. With specialized foam, carbon plates, and mesh to choose between, providing consistent, expert-level product advice across the company’s vast catalog is a top priority. PUMA set out to clean up its backend data systems and build an interactive, multilingual conversational agent to make expert recommendations accessible to all of its customers.
*Solution:* Crafting a high-end interactive experience led PUMA to a team of supporters from Google Cloud, NVIDIA, and LiveX. Together, these teams created PUMA’s AI-powered virtual assistant, Dylan, named after the company’s internal name for the iconic cat logo. PUMA deployed the conversational agent on Google Cloud with Google Kubernetes Engine (GKE) and Cloud Run, so the system could scale up dynamically during peak in-store hours or sales and scale down to conserve resources when stores closed. It also uses the flexibility of a multi-model approach to match AI model selection to different use cases. For example, it can use Gemma to resolve simpler inquiries with ultra-low latency, and route more complex product questions to NVIDIA Nemotron. Together with LiveX, PUMA seamlessly ingested its entire product catalog, creating a truly unified, multilingual shopping experience.
*Results:* Since launching Dylan on the retail floor, PUMA has seen the assistant autonomously resolve 75% of customer inquiries. Store associates are grateful to have these cutting-edge AI tools available that now free their time to address more complex guest interactions. By supporting more than 100 languages, Dylan successfully removes language barriers for international shoppers and helps any shopper navigate PUMA's catalog. Instead of struggling with specialized technical terms, customers interact with Dylan through natural, conversational questions and responses that guide them from discovery to trying on a pair of shoes to checking out with a staff member.
*Key takeaways and highlights from our interview with Ivan Dashkov, Head of Emerging Marketing Technology at PUMA:*
→ “Our store associates hear about AI all the time, and I love that they’re getting to play with this new technology while it’s helping them do their jobs. People come into our stores speaking different languages, and sometimes our team doesn’t know how to help those people. Now they can send them to Dylan.”
→ “If we didn't have partners like Google Cloud, NVIDIA, and LiveX, I don't think we'd have this project. I don't think we’d have our AI creator project. The success of these projects is because of these partnerships.”
→ “This world is moving so fast and having partnerships we can leverage with NVIDIA and Google to deliver best-in-class consumer experiences is such a huge advantage for us.”
*Google Cloud products used:* Cloud Run, Gemini platform, Google Kubernetes Engine
*Learn more:*
→ Gemini platform: Innovate, build, and deploy enterprise ready agents and apps https://cloud.google.com/products/gemini-enterprise-agent-platform
→ Cloud Run: Build apps or websites quickly on a fully managed platform https://cloud.google.com/run
→ Kubernetes, evolved: The foundation for platform builders https://cloud.google.com/kubernetes-engine
*Executive summary:* PUMA teamed up with Google Cloud, NVIDIA, and LiveX to deploy Dylan, an innovative, AI-powered virtual assistant that delivers expert product advice to shoppers. Built on Google Kubernetes Engine and Cloud Run, the system dynamically scales to handle peak retail traffic. By taking an agile multi-model approach, PUMA uses Gemma for quick inquiries and routes detailed catalog queries to NVIDIA Nemotron. Dylan securely ingests PUMA's extensive product catalog, supporting over 100 languages to eliminate shopping barriers for international guests. The virtual assistant now autonomously resolves 75% of in-store customer inquiries, allowing store associates to prioritize complex interactions and providing shoppers with a seamless, conversational journey from product discovery to checkout time.
*Challenge:* As one of the world's premier sports brands, PUMA prides itself on delivering high-end experiences to both everyday consumers and globally-ranked athletes. Even common products like running shoes have become sophisticated tools. With specialized foam, carbon plates, and mesh to choose between, providing consistent, expert-level product advice across the company’s vast catalog is a top priority. PUMA set out to clean up its backend data systems and build an interactive, multilingual conversational agent to make expert recommendations accessible to all of its customers.
*Solution:* Crafting a high-end interactive experience led PUMA to a team of supporters from Google Cloud, NVIDIA, and LiveX. Together, these teams created PUMA’s AI-powered virtual assistant, Dylan, named after the company’s internal name for the iconic cat logo. PUMA deployed the conversational agent on Google Cloud with Google Kubernetes Engine (GKE) and Cloud Run, so the system could scale up dynamically during peak in-store hours or sales and scale down to conserve resources when stores closed. It also uses the flexibility of a multi-model approach to match AI model selection to different use cases. For example, it can use Gemma to resolve simpler inquiries with ultra-low latency, and route more complex product questions to NVIDIA Nemotron. Together with LiveX, PUMA seamlessly ingested its entire product catalog, creating a truly unified, multilingual shopping experience.
*Results:* Since launching Dylan on the retail floor, PUMA has seen the assistant autonomously resolve 75% of customer inquiries. Store associates are grateful to have these cutting-edge AI tools available that now free their time to address more complex guest interactions. By supporting more than 100 languages, Dylan successfully removes language barriers for international shoppers and helps any shopper navigate PUMA's catalog. Instead of struggling with specialized technical terms, customers interact with Dylan through natural, conversational questions and responses that guide them from discovery to trying on a pair of shoes to checking out with a staff member.
*Key takeaways and highlights from our interview with Ivan Dashkov, Head of Emerging Marketing Technology at PUMA:*
→ “Our store associates hear about AI all the time, and I love that they’re getting to play with this new technology while it’s helping them do their jobs. People come into our stores speaking different languages, and sometimes our team doesn’t know how to help those people. Now they can send them to Dylan.”
→ “If we didn't have partners like Google Cloud, NVIDIA, and LiveX, I don't think we'd have this project. I don't think we’d have our AI creator project. The success of these projects is because of these partnerships.”
→ “This world is moving so fast and having partnerships we can leverage with NVIDIA and Google to deliver best-in-class consumer experiences is such a huge advantage for us.”
*Google Cloud products used:* Cloud Run, Gemini platform, Google Kubernetes Engine
*Learn more:*
→ Gemini platform: Innovate, build, and deploy enterprise ready agents and apps https://cloud.google.com/products/gemini-enterprise-agent-platform
→ Cloud Run: Build apps or websites quickly on a fully managed platform https://cloud.google.com/run
→ Kubernetes, evolved: The foundation for platform builders https://cloud.google.com/kubernetes-engine
Google Cloud 다른 영상
Introducing the Google Cloud Professional Agentic Architect Certification | Get certified today
조회수 1
Safely Delegate Real Work to AI on Google Cloud | Barndoor
조회수 14
How Starling Bank's AI serves 5M customers and boosts canceled scam payments 300%
조회수 42
Build self-evolving AI agents with Reindeer on Google Cloud
조회수 202
Application awareness on Cloud Interconnect demo
조회수 565
Build stronger customer experience outcomes with voice, text, and speech AI
조회수 326
New Way Now: OTTO brings real-time voice shopping to its app with Google Cloud
조회수 580
New Way Now: Bosch scales Agentic AI with Gemini on Google Cloud
조회수 645