AI Agent sub process not calling REST connector as tool

Camunda 8 Topics
1

I am trying to create one small POC for Agentic AI process creation. I have configured AI agent subprocess with REST connector which is pointing to local REST endpoint.
This rest endpoint is running and working from postman.
I have configured local ollama and that is also running with model present that is configured here.

I see token stuck in AI Agent subprocess and not moving forward.
BPMN is attached.
agenticAI.bpmn (13.0 KB)

2

This looks like a common AI Agent subprocess issue related to connector runtime, tool feedback loop configuration, missing tool documentation, or output mapping to toolCallResult. I found the following relevant resources:

Does this help? If not, can anyone from the community jump in? :waving_hand:

:light_bulb: Hints: Use the Ask AI feature in Camunda’s documentation to chat with AI and get fast help. Report bugs and features in Camuda’s GitHub issue tracker. Trust the process. :robot:

3

Still no luck to identify issue.

4

@Niall @cpbpm your help would be appreciated

5

What AI provider you selected to point to Ollama? You should select OpenAI as a provider and point your local Ollama endpoint.

Did you check Niali’s youtube video: https://www.youtube.com/watch?v=E0BaMLL5Uf4&t=5s

6

@cpbpm Yes, I followed his video tutorial, but it has Amazon Bedrock configuration, however few Camunda pages suggests OpenAI compatible for ollama local.

Ref

OpenAI has more props to configure like, api key, org id, project id, which I don’t see available for ollama local setup.
Pl. suggest on this observations.

7

Give me few days, let me try out and update you with relevant configuration.

8

Just FYI, I am not focusing on end to end agentic ai, just only expecting REST connector should run in AI Sub process context and will verify result on operate.
Then I can enhance this further.

Thanks

9

Hi

I have successfully run an agent subprocess using local ollama running google gemma… in addition I include a rest connector tool and all runs fine, but slow given its all running on my local laptop…

I suggest diagnosing whats going on inside the llm first, eg have a look at ollama logs to see that its actually doing something…

Regarfs

Rob

10

Sure, if you could review the bpmn attached, to check if configuration looks correct or anything missing, it would be good.

Also, I’m using Ollama llama3.2 model on local, FYI.

11

@Webcyberrob I got this issue resolved. I was using old connector runtime. Running using latest 8.9, and its working.
Thanks for your help !!!