How to Build AI Products Across B2B and B2C: Stories from Uber, Microsoft & Instacart
summary
Describes building a personalized promotions ML product for Instacart (piloted with ~10 brand partners, 2–3 deferred) and a multimedia-powered product ads system for Bing/Edge that extracts products from images and drives millions in incremental ad revenue. Covers concrete steps: data permissions and legal docs, model metrics like precision/recall and CTR/CPC, offline validation then A/B testing at 5% rollout, and post-launch monitoring with alerts for model drift and retraining to maintain performance.
AI-written summary · about
chapters (4)
related talks
23mDelivering business success with AI
42mAI-powered products and product-powered AI
24mAI-Powered Transformation in Fashion E-Commerce
31mWant to adopt AI in your business: Good luck!
20mBeyond Relevance: Using LLMs to Transform E-commerce Search into Product Discovery Engines
41mAI + Machine Learning for .NET Developers
26mHow to develop useful and unbiased AI
21mUsing AI to Scale Product Growth