# How to Build AI Agents That Don't Start Over When They Fail

DevFeed: [How to Build AI Agents That Don't Start Over When They Fail](<https://devfeed.tech/articles/how-to-build-ai-agents-that-don-t-start-over-when-they-fail-17911.md>)

Original publisher: [Read original article](<https://newsletter.systemdesign.one/p/durable-ai-agents>)

Author: Neo Kim

Published: 2026-08-20T17:26:04Z

Content type: tutorial

Language: en

Sources: [System Design Newsletter](<https://devfeed.tech/sources/system-design-newsletter.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [build](<https://devfeed.tech/tags/build.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [workflow](<https://devfeed.tech/tags/workflow.md>)

## AI overview

This tutorial explains why long-running AI agent workflows need durable execution. It describes how tracking completed steps can let a workflow resume after failures, retries, waits for human approval, or loss of the process running it, using Inngest as a case study.

## Source excerpt

#170: The Infrastructure AI Agents Actually Need