# Building a small Erlang service around an AI-generated MVP - what should I check first?

**URL:** <https://erlangforums.com/t/building-a-small-erlang-service-around-an-ai-generated-mvp-what-should-i-check-first/6107>\
**Category:** Questions / Help\
**Tags:** community\
**Created:** [November 5, 2026, 8:11am UTC](https://erlangforums.com/t/building-a-small-erlang-service-around-an-ai-generated-mvp-what-should-i-check-first/6107 "2026-11-05T08:11:46Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![nicholasjohnson](https://erlangforums.com/letter_avatar_proxy/v4/letter/n/8baadc/32.png) [@nicholasjohnson](https://erlangforums.com/u/nicholasjohnson)\
**Post date:** [November 5, 2026, 8:11am UTC](https://erlangforums.com/t/building-a-small-erlang-service-around-an-ai-generated-mvp-what-should-i-check-first/6107/1 "2026-11-05T08:11:46Z")

</div>

I have been learning Erlang on and off for a while and want to use it for a small side project rather than going back to something more familiar. I came across third-party information online about building an MVP quickly with AI app builder tools, and the idea of generating a rough starting point appeals to me since I am not strong on frontend work.

What I am unsure about is how much of a generated MVP actually fits an Erlang backend. I would rather keep the core logic in Erlang processes and treat any generated code as scaffolding I rewrite, but I do not know where the seams usually break. Supervisors and message passing seem easy to get wrong if the structure comes from a tool that assumes a typical request-response stack.

Has anyone here started a project this way and kept the BEAM side clean? I am mainly curious what to inspect first and what to throw away outright. Any input from people who have done it would help.