Building Deep Agents with Scale AI and Temporal

Most AI interactions today are transactional: you prompt the bot, it answers, and the interaction ends. But the future of enterprise AI is long-running agents that work autonomously in the background for weeks or months, asking for human input only on key decisions. Scale AI recently open-sourced Agentex to make this possible, partnering with Temporal to provide the durable execution layer necessary for reliability. By combining Agentex's orchestration with Temporal's ability to survive failures and wait indefinitely, developers can now build agents that manage extended workflows--like construction procurement--without the reliability woes of standard scripts. In this webinar, experts from Scale AI and Temporal sit down to discuss the architecture behind these self-driving agents. We explore how to flip the paradigm from "humans prompting AI" to "AI prompting humans," followed by a deep dive into the code. Key takeaways: - Understand the architectural shift required to move from stateless chatbots to persistent, event-driven agents that can run for months. - Learn how to use Agentex with Temporal to build workflows that handle restarts, crashes, and long-dormant periods without losing state or context. - See a live demo of a "Procurement Agent" that autonomously tracks shipments, manages schedules, and proactively escalates edge cases to humans. - Explore advanced patterns like automatic conversation summarization and feedback loops where agents get smarter with every human interaction. --- Temporal is the simple, scalable, open source way to write and run reliable cloud applications. Learn more Blog: https://temporal.io/blog How Temporal Works: https://temporal.io/how-temporal-works Community Slack: https://temporal.io/slack Developer resources Docs: https://docs.temporal.io Courses: https://learn.temporal.io/courses Support forum: https://community.temporal.io