Skip to main content
Question

Title: AI Automation Agents – Roadmap for Additional LLM Providers?

  • Oct 28, 2026 6:46 PM UTC
  • 2 replies
  • 20 views
  • Translate All

elijahr83
Semi-Pro I
Forum|alt.badge.img+1

Hey everyone!

I've been exploring the AI Automation agent functionality in 2026 R2 and was curious if there's been any additional discussion around expanding the supported LLM providers.

Currently, we have OpenAI, Anthropic, Azure, and AWS Bedrock available through Agent LLM Connections. I saw the community poll earlier this year regarding potential support for additional providers like Google Gemini, Grok, Mistral, and others.

Has Acumatica shared any updates on whether additional providers are planned for an upcoming release, or if there's a tentative timeline for expanding the available options?

I'm particularly interested in whether we'll eventually see more flexibility in selecting models from additional providers when configuring agents.

Would love to hear if anyone has any insight into what's on the roadmap!

2 replies

TimRodman
Pro III
Forum|alt.badge.img+1
  • Pro III
  • Oct 28, 2026 7:28 PM UTC

​@elijahr83 - no updates that I’m aware of. For me personally, I’m often tempted to geek out on the models (which one is better, etc.), but I keep reminding myself of the 80/20 rule in ERP:

Attack the 80% in ERP and leave the 20% for spreadsheets and email.

With that mindset, I’m just focused on using an LLM (any LLM) in ERP workflows and proving out those use cases before getting into the 20% of which model is better for which task.

Translate

elijahr83
Semi-Pro I
Forum|alt.badge.img+1
  • Author
  • Semi-Pro I
  • Oct 28, 2026 7:41 PM UTC

​@TimRodman Geeking out is exactly how I would describe what I’ve been going through! haha

However, I definitely agree that proving the practical use cases should be our priority for the time being. I think where I’m coming from is more about flexibility for prospects and customers; Giving them the ability to map out their AI connectors for future use. 

But I completely agree that getting the use cases right is ultimately what matters most!

Translate