Rethinking the future of digital transactions
summary
Enterprise agents run unattended coding for seven hours (Rakuten), replace whole teams (a 50-person hospital supply group), and generate millions of per-user apps that each need a cheap DB and checkpoints, forcing Neon/Databricks-style common storage and governance. Stripe ships an Agent Commerce Protocol used for one-click purchases from Etsy in ChatGPT, and internal agents cut a 10-SDR workload to roughly $1,000/year to run, shifting pricing, token billing, and read-to-write permission models. Vercel demos and MCP servers appear throughout.
AI-written summary · about
chapters (12)
- 0:00Agentic applications in enterprise
- 1:04Agentic use cases in industry
- 3:29AI agents impact
- 5:16Internal AI automation success
- 8:45AI transformation key elements
- 10:54AI governance challenges
- 15:28AI agents data needs
- 20:07Stripe’s agent commerce protocol
- 24:59Agent-driven commerce future
- 27:05Pricing for agentic systems
- 31:39AI adoption habits
- 33:44Closing and Q&A
related talks
56mOpening Keynote
25mAgents are not Apps: Stop building interfaces, Start building coworkers
43mThe AgentVerse - Build fast, Scale smarter
21mBuilding Agents (the hard parts!)
15mConquering Agent Chaos
54mKeynote
28mFrom Demo to Durable: Building AI Companies That Actually Scale
15mThe Responsible Path to Hyperproductivity with AI Agents