Opentext migration – can ai actually convert legacy process docs to working workflows?

We’ve got 500+ pages of legacy OpenText documentation and zero current expertise. Leadership wants AI to auto-convert these to Camunda/n8n flows. Anyone tried automated process mining from old BPM docs? How accurate were the generated workflows? Did you still need manual cleanup?

Used Latenode’s AI Copilot on 200+ page IBM BPM docs. Generated 80% of core workflows automatically. The key was feeding it our old test cases for validation. Still needed manual tweaks but cut migration time by half. Try their doc-to-workflow conversion: https://latenode.com

tried uipath doc understanding. worked okay 4 simple flows but failed on exceptions. had 2 rewrite all error handling. maybe combine with RPA?

Process mining tools like Celonis work better with system logs than documentation. For docs, we found manual mapping unavoidable. Created a hybrid approach where junior staff handles AI-generated base workflows while seniors validate against original SLAs. Still time-consuming but more reliable than full automation.