Slide 1

Slide 1 text

Gaétan Rougevin-Baville CEO Cadillacs, Condos & Cardigans: Leveraging AI for Consumer Merchandising Real-World AI Implementation in Visual Commerce

Slide 2

Slide 2 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Processing millions of images monthly

Slide 3

Slide 3 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Processing millions of images monthly

Slide 4

Slide 4 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Expectations Time AI: Hype vs Reality SECTION 1 The reality of AI in production

Slide 5

Slide 5 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising & / or Σ INDIVIDUAL IMPACT PROCESS OR PRODUCT IMPACT AI: Achieving B2B Impact SECTION 1 The reality of AI in production

Slide 6

Slide 6 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising CREDIBILITY USABILITY RELIABILITY B2B Industry Expectations (<>B2C) SECTION 1 The reality of AI in production

Slide 7

Slide 7 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising ● Integrated guidelines ● Integrated workflow ● Quality check ● No training required ● Plug & play ● Faster time to market ● Higher conversion ● Cost savings Spot on! Ready to take image. CarCutter user process 1 2 3 SECTION 2 Vertical-Specific Solutions & Results

Slide 8

Slide 8 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising ● Integrated guidelines ● Integrated workflow ● Quality check ● No training required ● Plug & play ● Faster time to market ● Higher conversion ● Cost savings Spot on! Ready to take image. 1 2 3 ProperShot user process SECTION 2 Vertical-Specific Solutions & Results

Slide 9

Slide 9 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising “Years to build a brand. Seconds to destroy it.” B2B Adoption SECTION 2 Vertical-Specific Solutions & Results

Slide 10

Slide 10 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising SECTION 3 Technical insights and lessons learnt B2B Adoption

Slide 11

Slide 11 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Input Image Inpaint ControlNet Foundation Real-Estate Model Off-the- shelf Canny ControlNet Layout ControlNet Upscale, refine and post-process with diffusion-based refiner Stylize eg: “traditional” Perform realistic integration Copy/paste furniture into input image Foundation Real-Estate Model Obtain furnished room with same layout Layout ControlNet Foundation Real-Estate Model SECTION 3 Technical insights and lessons learnt Fast, agile deployment

Slide 12

Slide 12 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Fast, agile deployment SECTION 3 Technical insights and lessons learnt

Slide 13

Slide 13 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Ends Users Verticalized AI products Platform Verticalized and proprietary data (Input media, AI output, annotated data model performance metrics, logs, etc.) 50% reduction inference time Human in the Loop Media processing Platform Benefits SECTION 3 Technical insights and lessons learnt

Slide 14

Slide 14 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Key learning points CREDIBILITY Value of vertical specific data & experts USABILITY Importance of workflow integration & UX RELIABILITY Balance of automation and human oversight SECTION 3 Technical insights and lessons learnt

Slide 15

Slide 15 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Future of AI in merchandising SECTION 3 Technical insights and lessons learnt

Slide 16

Slide 16 text

Cadillacs, Condos and Cardigans: Leveraging AI for Consumer Merchandising Gaétan Rougevin-Baville [email protected] Get in touch: