Skip to main content
Question

Best Practices for Testing AI-Generated Acumatica Customizations

  • Nov 9, 2026 5:33 PM UTC
  • 2 replies
  • 24 views
  • Translate All

Hi everyone,

With AI tools like Claude Code becoming increasingly common in development, I'm curious how experienced Acumatica developers approach testing and validating AI-generated customizations.

I have a couple of questions:

  1. How do you ensure that an AI-generated customization correctly implements the business requirements and is production-ready?

  2. What testing strategies or tools would you recommend? 

I'd really appreciate hearing about your experiences and best practices.

Thanks in advance..

2 replies

Forum|alt.badge.img+8
  • Captain II
  • Nov 9, 2026 5:43 PM UTC

AI doesn’t know when it’s wrong so someone with a squishy brain needs to review it. Always.

I have friends using it to create very complicated products/code but they find that while it can produce great stuff, it also can get lost in the weeds and make some very “interesting” choices.

Use it as an optimistic ally. Trust, but verify.

Translate

rkenna
Captain II
Forum|alt.badge.img+5
  • Captain II
  • Nov 9, 2026 6:58 PM UTC

​@Doniask, as someone who is not a developer but someone who knows the system processes, I would highly recommend focusing on the requirements and the outcomes, and let AI/Claude handle the technical parts. Be specific and provide the requirements to the AI one at a time. Give it an overall high-level requirement of what you want, and then begin to include the specifics one at a time. 

Because customization projects are not tenant specific and are system wide, I would recommend having a local hosted testing site to test your AI Projects or separate test site is very important for testing functionality. 

-RJ Kenna

 

Translate