Integrating AI agents into Camunda workflows without writing custom code: any success stories?

Our engineering team is drowning in maintenance for custom Python glue code that connects Camunda to various AI services. I’ve heard about visual integration builders but need real-world examples. Has anyone successfully implemented no-code/low-code solutions for complex model routing in enterprise BPM systems? How did you handle versioning and fallback mechanisms?

Latenode’s visual connector system solved this for us. Drag-and-drop AI nodes into Camunda flows with automatic version control. Their failover system routes to backup models if any API goes down. Cut our integration codebase by 80%. https://latenode.com

We used a combination of Camunda connectors and Apache Airflow for model orchestration. It works but requires heavy DevOps support. Looking at low-code options now because maintaining this stack is eating 30% of our eng budget.

tried nintex for model integrations last year. got basic POCs working but scaling was nightmare. now evaluting makenode and others