Build AI agents for e-commerce with ADK + Vector Search

This video discusses creating AI agents for e-commerce using ADK and Vector Search to address challenges like keyword/multimodal searches and item recommendations. It highlights advanced Vector Search practices (multimodal, hybrid, task-type embedding) and demonstrates how AI agents with Vector Search improve ""generative"" recommendations, exemplified by a ""Shopper's Concierge"" demo. This demo, using ADK, Google Search grounding, query generation, and multimodal item curation, finds relevant items based on user intent, moving beyond simple text similarity search.

Resources:

Vertex AI Vector Search →https://goo.gle/3T5xxK5
Agent Development Kit→https://goo.gle/3RGrB9T
Shopper's Concierge demo video→https://goo.gle/4jRbMJb
Shopper's Concierge sample notebook→https://goo.gle/4kMkxot

Chapters:
0:00 - Intro
0:43 - Introduction to Vector Search & Advanced Practices (Kaz Sato)
1:18 - Typical RAG Scenario for E-commerce
2:56 - Challenges with Basic RAG in E-commerce
8:08 - Vector Search Advanced Practices
15:08 - Building the Shopper's Concierge AI Agent with ADK & Vector Search (Kaz Sato)
8:50 - Challenge: Smart Recommendations
9:50 - Solution: AI Agents + Vector Search
11:13 - Shopper's Concierge Demo
14:58 - Code Implementation Walkthrough
15:02 - Q&A and Resources
15:48 - Why ""Agent as a Tool"" vs. Sub-agent?
16:48 - Getting Started Resources

Subscribe to Google for Developers → https://goo.gle/developers

Speakers: Sita Lakshmi, Kaz Sato
Products Mentioned: Agent Developer Kit (ADK)