### Answer Yes, using more than 20 nodes in Flex Mode can be problematic. The documentation explicitly states: > "If you use flex mode for more than 20 nodes, performance might degrade and agent might have higher hallucination risk." [(1) ]() Specific issues you may encounter: - Performance degradation - Higher hallucination risk (the agent may generate responses not aligned with your instructions) Recommended solution: Split your flow into smaller components. The documentation recommends "splitting into smaller components" when you exceed 20 nodes. This approach lets you enable Flex Mode on individual components rather than the entire agent. When Flex Mode is enabled on a component, only that component's nodes are converted into a single prompt—the rest stays as standard conversation flow. So while it's not impossible to use 20+ nodes, you should expect reliability issues. Breaking your flow into logical component groups under 20 nodes each is the best practice. , [(2) ]()

​ 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) ​

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

### Answer --- Based on the available sources, I don't have explicit documentation stating whether the LLM automatically stays in a Conversation Node until a transition is triggered, or whether you need a self-loop edge for repeated FAQ questions. However, here's what I can share from related information: How Conversation Flow transitions work: Nodes have edges with transition conditions that determine when to move to another node. [(1) ]() The LLM evaluates these conditions based on user input. If no transition condition is met, the behavior depends on your flow configuration. Community reports on looping issues: There have been reports of conversation flows experiencing looping problems, particularly with Global Nodes and split nodes. [(2) ]() Some users have needed to add transition fine-tune examples to help the model understand when to stay vs. transition. Recommendation to reduce node count: The Prompt Engineering Guide suggests using Conversation Flow when you have "5+ different functions/tools or multiple decision branches." [(3) ]() This implies that well-structured prompts within nodes can handle multiple interactions. What you can try: - Write clear instructions in your FAQ node's prompt specifying it should continue answering questions until a specific exit condition is met - If you find the agent transitioning unexpectedly, adding a self-loop edge with a condition like "user asks another FAQ question" may help stabilize behavior I don't have definitive documentation on whether a self-loop is required. For a more authoritative answer, check out these potentially useful sources: - [Prompt Engineering Guide](

/prompt-engineering-guide) - [Documentation]() - [Hire certified Retell partner]() Would you like a general answer based on my knowledge instead?

​ To learn more, see the sources I considered: [[1] List Conversation Flows](https://docs.retellai.com/api-references/list-conversation-flows) [[2] Conversation Flow Node Looping](https://discord.com/channels/1186424376300679209/1372978630841602078) [[3] Prompt Engineering Guide](https://docs.retellai.com/build/prompt-engineering-guide) ​

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