AI at Scale in the Enterprise
summary
Enterprise moves from prototype to production by proving models at a factory then centralizing deployment to scale across sites, using AI ladder practices: curate diverse data, iterate training on GPU servers, verify fairness and explainability, and automate rollouts. A concrete build uses IBM Maximo Visual Inspection fed by iOS cameras into dashboards that trigger robotic actions; outcomes report reduced manual inspection and repeatable POCs. Architecture emphasizes metadata, single namespace storage, and edge inference for low latency.
AI-written summary · about
chapters (12)
- 0:00Intro
- 0:41How Customers Do Data Analytics Traditionally
- 1:38Enterprise Analytics Modernization: From Data to Actions
- 2:38Common Patterns of Analytics Business Problems
- 3:24Three broad categories of Use Cases
- 6:37The AI Ladder
- 7:47Data is Prerequisite to AI
- 9:16Enterprise Data Pipeline for AI
- 11:44AI Model Development Workflow
- 15:44Common AI Data Considerations
- 18:14Typical AI Inferencing Scenarios
- 23:44A Packaged Goods Quality Inspection Example
related talks
29mFrom Experiment to Enterprise: Architecting AI for Stability and Scale
26mAI for enterprise teams
12mCreating a fair and trusted AI enterprise
56mPanel: Scaling AI successfully
29mAI Operationalization: Where the AI Rubber Hits the Road for the Enterprise
24mAI Implementation: Navigating the Enterprise Lifecycle for Technical Leaders
19mScaling & Industrialization of AI
16mUsing AI and Computer Technologies to Enhance IT Infrastructure