Announcing New AI Capabilities
summary
Temporal ships durable orchestration for agentic apps: a six‑step returns agent that calls LLMs, checks order history, inventory, waits durably for human approval, then issues a refund via a mock Stripe API and sends email without holding compute. Features shown include workflow streams (public preview) for UI event delivery, plugins for OpenAI, Google ADK and Anthropic to run different agent SDKs, S3 large‑payload storage, and priority and fairness keys to control concurrency and rate‑limit retries.
AI-written summary · about
chapters (13)
- 0:00Introduction and speakers
- 1:13Agentic AI value and growth
- 4:02Temporal workflow primer
- 9:42Demo product return agent
- 14:50Introducing Draft application
- 18:03Context preparation pipeline
- 24:47Handling rate limits and fairness
- 27:20Agentic side architecture
- 31:03Temporal AI plugins
- 33:57Workflow streams for UIs
- 36:05Draft demo recap
- 40:22Observability and integrations
- 41:47Closing and future
related talks
34mThe Hard Truth About AI Agents in Production
33mFrom Prototype to Production: Making Agentic AI Systems Reliable
11mCreate AI Agents That Work Independently
27mAI-Powered ETL Pipeline Orchestration: Multi-Agent Systems in the Era of Generative AI
31mBuilding Agentic AI Platforms
30mAI Agents Unchained
29mWhat exactly are AI Agents?
29mLeveraging AI to Enhance DevEx & Operational Efficiency