How to run a consistent end-to-end workflow where some tasks are simple, others simple agents, others human in loop, and others more complex agentic AI tasks (with complexity embedded) - but always having one macro end-to-end journey being managed
Hey @Maldini3105 thanks for posting! Want to make sure I understand how to help… are you asking for workflow design guidance using Orkes Conductor? Or something else I’m missing?
I was asking what is the control handoff between a master workflow and within same a task that is an agentic AI task with its own internal complexities - then return to master workflow
Hi @Maldini3105 ! There’s no particular uniqueness task-wise or workflow-wise with agentic tasks/workflows vs other tasks/workflows - they are interchangeable as far as the workflow goes, so you can kick off an agentic task or workflow/subworkflow sync or async just like you would any other! All tasks, regardless of the work they do, fit within the workflow framework fairly equally. Does this answer your question? If not, perhaps you could provide some more context, i.e. how are you defining “agentic AI task” - are you referring to a custom task that does agentic work, or perhaps referring to one of our LLM tasks?