AI Agents in Practice
summary
Live demo builds an agent called Mrs GoTo that uses Claude 4.5, OpenAI transcribe, GitHub and Slack integrations, phone calls and email access, and schedules a weekly research job that runs autonomously and posts summaries. A separate scheduling agent ingests spreadsheets as JSON data documents, writes a reusable import script and a CRUD app so humans review suggestions, cutting time and improving decision quality (investment screening proves more consistent than prior human reviews). An HTTP requests tool plus Perplexity and OpenStreetMap produces maps and live bus times on demand.
AI-written summary · about
chapters (14)
- 0:00Intro
- 0:21The future of work
- 1:02What is an AI agent?
- 2:32Demo: How to build an AI agent
- 11:40Code vs Agent vs Human
- 13:12Architecture of an agent platform
- 14:47Where to use agents
- 19:13Content is king
- 22:30Demo: Data documents & apps
- 24:13Content is king continued
- 27:11Agent design tips
- 33:20Cool thing: Agents making their own tools
- 36:45Wrap-up
- 39:35Outro
related talks
49mAI Agents: Collaboration in Real-Life Applications
45mAI agents in practice - beyond the prototype
01h 30mHow to build AI Agents: A step by step guide to creating AI agents
26mAI agents 101
01h 04mAgentic AI: Build a Multi-Agent Application with CrewAI
52mCut the Hype and Build Your Own AI Agents
41mGoing beyond copilot with AI Agents
25mWho Let the Bots Out? A Guide to Evaluating AI Agents