Hi everyone,
I’d like to share some hands-on experience configuring AI Studio in Acumatica, specifically for generating transaction summary directly inside the ERP.
Recently, we tested AI Studio in a real implementation and found that it can be very effective when set up properly – especially for reducing manual content work while keeping everything inside Acumatica.
Here is the approach we followed:
1. Enable AI Studio
AI Studio must first be enabled in Enable/Disable Features (CS100000). Once enabled, the AI Studio screens become available.
2. Create an LLM Connection
We configured an OpenAI connection in LLM Connections (ML201000).
From a cost and performance perspective, we found gpt-4o-mini to be a good balance for ERP use cases like descriptions and summaries.
3. Design LLM Prompts carefully
Prompts were created in LLM Prompts (ML202000) and bound directly to the transaction screen.
One practical tip that helped a lot:
we copied the full DAC Schema of InventoryItem and used ChatGPT/Gemini to help draft accurate instructions aligned with real Acumatica fields.
4. Test before rollout
Testing prompts with real data inside ML202000 helped catch issues early before users started using the “Generate Description” button.
We documented the full step-by-step configuration (with screenshots) here:
👉 https://s4.com.vn/en/how-to-configure-ai-studio-in-acumatica-step-by-step-guide/
Curious to hear from others:
-
What AI Studio use cases have you tried so far?
-
Are you using it for descriptions, summaries, or something more advanced?
Looking forward to learning from the community.