Keeping pets happy and healthy: Purina’s journey with genAI
summary
Purina prototypes a generative AI pet‑trainer chatbot using Amazon Bedrock, tying product data in Aurora/OpenSearch via RAG and a serverless API front end; experiments cover ~20 prompting styles and tune creativity, with guardrails for fact‑checking and hallucination detection. They correct product claims (replace "Fel d5" with Fel d1), keep consumer data in a VPC, and rely on human review. Purina also leverages connected hardware like the Petivity Smart Litter Box and reports 100,000 quality checks per day while feeding 46 million dogs and 68 million cats annually.
AI-written summary · about
chapters (9)
- 0:00Introduction and agenda
- 1:26Purina business overview
- 2:57Customer journey and vision
- 4:09Generative AI goals
- 4:46AI implementation strategies
- 14:37Amazon Bedrock features
- 18:57Reference architecture
- 22:40Risk and responsibility
- 23:04Closing remarks
related talks
47mOutsmarting the Genie: Smart Architecture Decisions for GenAI Apps
36mGENAI-Powered chatbot to increase team efficiency
43m3 Steps to Create Trustworthy Generative AI Products
48mHow to Make Smart Architecture Decisions when Building Gen AI Apps
27mBuilding Generative AI features
15mPractical Evolution of Data Science with GenAI
28mIs GenAI the technology revolution we need? How can we harness it at work?
28mImproving responses in generative AI applications