### Answer Here's a comparison based on the documentation: ## Setup 1: Flex Mode, No Global Nodes Positives: - Combines conversation flow structure with single prompt flexibility for dynamic task handling [(1) ]() - Agent can easily switch context between different tasks (every node essentially becomes accessible) - Can move to the proper task if the user completes multiple tasks at the same time - After switching context to another flow, the agent can resume the previous task without repeating already completed steps - Good for handling varied user behavior naturally Negatives: - Performance may degrade with more than 20 nodes, and hallucination risk increases - Only use Prompt edges—avoid Equation edges as LLM interprets equation conditions poorly - LLM might not always follow static text instructions - Node-level Knowledge Base is ignored; you must configure KB at agent level - No "Speak After Execution" or "Wait For Result" settings for functions—agent always waits for function completion ## Setup 2: Rigid Mode, All Global Nodes Positives: - Strict, error-proof sequencing when you need deterministic behavior [(2) ]() - Global nodes allow users to skip or jump between nodes, useful for inbound support cases without a rigid call structure [(3) ]() - You can add fine-tuning examples for global nodes to control when the agent should/shouldn't jump to them [(4) ]() Negatives: - Making every node global can create unpredictable jumping behavior if transitions aren't carefully defined - May require more edges and transition conditions to handle all scenarios - Less natural handling of multi-topic

​ To learn more, see the sources I considered: [[1] Flex Mode](https://docs.retellai.com/build/conversation-flow/flex-mode) [[2] Flex Mode, Advanced Warm Transfer, 60+ languages, and more.](https://www.retellai.com/changelog/flex-mode-advanced-warm-transfer-60-languages-and-more) [[3] Debug guide](https://docs.retellai.com/build/conversation-flow/debug-guide) [[4] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) ​

conversations compared to Flex Mode Recommendation: Flex Mode works best when customers jump between topics or have multiple requests in a single call [(2) ]() . Rigid mode with global nodes is better when you need strict sequencing but still want escape hatches for specific scenarios .

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.