---
title: "Discussion on \"Why AI Applications Need Background Workflows\" | Hashnode"
description: "Discussion on \"Why AI Applications Need Background Workflows\". why AI applications need background workflows for long-running tasks, retries, webhooks, AI agents, and reliable asynchronous processing with Inngest."
url: "https://hashnode.com/posts/why-ai-applications-need-background-workflows/6aa6f5fd898e776851d641d9"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Why AI Applications Need Background Workflows](https://blogbysameer.hashnode.dev/why-ai-applications-need-background-workflows?utm_source=hashnode&utm_medium=feed)

why AI applications need background workflows for long-running tasks, retries, webhooks, AI agents, and reliable asynchronous processing with Inngest.

Sep 13 · 16 min read · blogbysameer.hashnode.dev  
Read the article

-   [#llm](https://hashnode.com/tag/llm)
-   [#rag](https://hashnode.com/tag/rag)
-   [#llms](https://hashnode.com/tag/llms)
-   [#ai](https://hashnode.com/tag/ai)
-   [#vector-database](https://hashnode.com/tag/vector-database)
-   [#memory-management](https://hashnode.com/tag/memory-management)
-   [#background-jobs](https://hashnode.com/tag/background-jobs)
-   [#ai-tools](https://hashnode.com/tag/ai-tools)
-   [#chaicode](https://hashnode.com/tag/chaicode)
-   [#generative-ai](https://hashnode.com/tag/generative-ai)
-   [#programming-blogs](https://hashnode.com/tag/programming-blogs)
-   [#programming-tips](https://hashnode.com/tag/programming-tips)
-   [#project-management](https://hashnode.com/tag/project-management)

## Written by

[Sameer Selokar](https://hashnode.com/@sameers9823)

@sameers9823 · Mumbai, India

  
[View profile](https://hashnode.com/@sameers9823)

Software Developer | Building Scalable Applications, Exploring AI, and Sharing Practical Tech Insights.

## Responses 1

Markdown supported · ⌘↵ to post

[Rulestack](https://hashnode.com/@rulestack)

[Sep 14](https://hashnode.com/posts/why-ai-applications-need-background-workflows/6aa6f5fd898e776851d641d9/comment/6aa81aec03a71de244e2af41)

Our unattended agent jobs run as cron workflows on GitHub Actions, where a mid-run failure leaves only whatever the command had already written to disk and no step for the next fire to resume from. The runs have been finishing in under three minutes, so I have not yet had to reach for the step-level tracking you describe.
