Actually shipping agents and not just talking about it
summary
Method: ask employees "what do you hate most about your job" and automate the boring, repeatable steps. Example: a lead processing agent uses LinkedIn and web scraping tools to classify contact forms that once took 50 minutes each, saving hundreds of days of work; an abuse agent renders pages, queries project metadata and returns a confidence score, cutting time per case ~59%; a Slack BI agent uses a Snowflake adapter to turn natural questions into SQL and charts. Versell reports ~400M deployments per year and open sources these workflows.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 3:27Versel and agent frameworks
- 4:28AI agents and potential
- 7:57Agent workflows basics
- 9:46Why agent projects fail
- 11:33Finding agent opportunities
- 13:16Lead processing agent
- 17:56Abuse detection agent
- 22:03Business intelligence agent
- 26:19Building your first agent
- 27:45Q&A
- 28:08DevOps automation with agents
- 34:59Security risks and prompt injection
- 37:49Software businesses and agents
- 40:40Agents and engineer laziness
- 41:21Abuse agent deep dive
- 43:51Monitoring agent tools
- 45:59Agents and human roles
- 47:46Junior engineers and AI
- 50:15Q&A wrap-up
related talks
36mAI Agents in Automation | 3 Real Production Cases with ROI
19mYour Biggest User Isn't Human: Designing DevTools for AI Agents
45mAI agents in practice - beyond the prototype
30mWhy Agents Don’t Ship: Closing the Production Gap for Real Business Impact
34mAI Agents for Performance: Ship Faster, Pay Less
16mMulti-Agent Intelligence for Next-Gen Insights
40mAI Agents in Practice
33mAgent Experience