Black Eyewear: Gemini-powered frame agent with Vertex AI Vision for personalized eyewear recommendations

Black Eyewear uses Google Cloud to extend its in-store optical consultation experience online. The company labeled thousands of images with Vertex AI Vision and trained multiple specialist models to capture nuanced facial-structure and frame-geometry judgments. It built a web-based frame agent powered by Gemini Flash to converse with shoppers and guide product exploration, and integrated virtual try-on using Nano Banana 2 orchestrated through Gemini Flash.

Organization
Black Eyewear
Industry
Retail
Published
June 2026

Reported outcomes

+20%

sales growthRevenue & growth

10 minutes or lessconsultation time60 minutesconsultation time4 secondstry-on generation time

Strategic outcomes

New product / capabilityLaunched AI-powered frame recommendation agentCustomer experience & trustImproved shopper confidence in frame selectionSpeed & agilityAccelerated optical consultations

Catalog median for revenue & growth deployments: +34% across 150 reported metrics. Compare benchmarks →

Primary read

Use case focus

Showing 3 of 4

  • 1Customer Experience
  • 2Personalization
  • 3Visual Search
  • Scale highly personalized optical consultations beyond the founder's in-store expertise.
  • Preserve nuanced taste and expert judgment as commerce shifts online.
  • Improve customer confidence while speeding up frame selection.
  • The team labeled thousands of images with Vertex AI Vision and trained multiple specialist models on curated face and frame data.
  • It built a web-based frame agent powered by Gemini Flash to interpret style cues and recommend frames.
  • It integrated virtual try-on image generation using Nano Banana 2 orchestrated through Gemini Flash.
  • Reported 20% sales growth.
  • Reduced in-store consultations from about an hour to ten minutes or less.
  • Customers felt more confident, more adventurous, and better prepared.
Architecture

Black Eyewear labeled thousands of images with Vertex AI Vision to train multiple specialist models that capture nuanced face and frame judgments. A web-based frame agent powered by Gemini Flash uses conversational cues and visual analysis to guide recommendations, and virtual try-on is generated with Nano Banana 2 orchestrated through Gemini Flash.

Sources & evidence1
Groundedness: 5/5Type: Customer StoryPublished: Jun 6, 2026Publisher: Google CloudEvidence: PrimaryConfidence: High

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Was this useful?