Beyond Relevance: Using LLMs to Transform E-commerce Search into Product Discovery Engines
summary
Instacart uses large language models as a merchandising engine that generates complementary, substitute and bundle carousels layered under exact-match search, launched late 2023 and expanded in 2024; injecting catalog and engagement signals shifts outputs from generic to contextual recommendations and yields 30%+ conversion lift on targeted queries, increases basket size and ad revenue, handles 14 million active users and billions of searches by serving dynamic personalized carousels in real time.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:02Search vs shopping differences
- 3:09Limitations of traditional search
- 5:20LLMs as merchandising engines
- 8:59Enhancing discovery with LLMs
- 10:45Challenges in conversational search
- 12:30Merchandising beyond exact matches
- 14:39Domain signals in LLM recommendations
- 16:26Business impact of AI merchandising
- 18:12Key learnings from LLM merchandising
- 19:40Future of dynamic merchandising
- 20:00Closing remarks
related talks
21mTransforming search and discovery using LLMs
22mMachine Learning: The Engine Behind Modern E-commerce Search
25mTransforming E-Commerce with LLMs: 3 Real-World Use Cases
51mQuery Understanding With Large Language Models: Techniques and...
47mModern Approaches to Product Matching Using LLMs
01h 30mUsing LLMs to improve your Search Engine
39mAI-Powered Search in Modern E-Commerce
22mAI-Powered Personalized Discovery with Graph Search and VLM-Enhanced Embeddings