### Answer A subagent node is used for conversations where the agent needs to talk with users while being able to call tools or functions during the dialogue. [(1) ]() It's designed for scenarios where the LLM should decide whether and when to use a tool based on conversation context. ## How It Works When a subagent node has tools or functions attached, the LLM receives both the node instruction and the list of available tools. During the conversation, the LLM determines when a tool should be called based on context, extracts the required parameters, and invokes it while maintaining the dialogue with the user. Key characteristics: - Multiple tools or functions can be added to a single subagent node - The agent can continue talking while a tool or function executes - Tool or function results are available to the LLM for generating follow-up responses ## Instructions Subagent nodes only support

Prompt

instructions—`Static Sentence` is not supported. You should write the instruction to define the task, what information the agent should gather, and when it should use the available tools. Example instruction:

Help the user check their order status. If the user provides an order number,
use the available order lookup tool to retrieve the latest status.

## Subagent Node vs Function Node Here's when to use each: - **Function Node**: Deterministic execution on node entry, one tool per node, not intended for dialogue, best for always-execute actions - **Subagent Node**: LLM-driven execution when appropriate, multiple tools per node, full dialogue with tools available, best for context-dependent actions during dialogue Use function nodes when you want guaranteed execution every time the flow reaches that step. Use subagent nodes when the agent should decide whether and when to call a tool based on what the user says. ## Available Tool Types You can attach various tool types to subagent nodes, including: - Custom Function - Code Tool - Check Calendar Availability - Book Appointment - End Call - Transfer Call - Transfer Agent - Press Digit - Send SMS - Extract Dynamic Variable - MCP Tool ## Important Note As of April 18, 2026, the

tools

and

tool_ids

fields on conversation nodes are deprecated in favor of the subagent node type. [(2) ]() Existing conversation nodes with tools will be automatically migrated to

type: "subagent"

.

​ To learn more, see the sources I considered: [[1] Subagent Node](https://docs.retellai.com/build/conversation-flow/subagent-node) [[2] Deprecating on 04/18/2026](https://docs.retellai.com/deprecation-notice/2026/04-18_conversation_node_tools) ​

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

### Answer Based on the available sources, there are no deprecation notices for multi-prompt agents (also called multi-state prompts) [(1) ]() . Multi-prompt agents remain a supported architecture option alongside single-prompt agents and conversation flow agents [(1) ]() [(2) ]() . Multi-prompt agents continue to be recommended for specific use cases where you need to organize conversations into a structured tree of states, each with its own focused prompt and behavior [(1) ]() . The platform changelog shows ongoing support and updates for multi-prompt functionality, including features like state transitions and edges [(3) ]() [(4) ]() . If you're concerned about future-proofing your implementation, conversation flow agents offer an alternative with more granular control through a node-based visual editor [(2) ]() . However, multi-prompt agents are not deprecated and remain fully functional for structuring complex conversations with multiple states [(1) ]() [(5) ]() .

​ To learn more, see the sources I considered: [[1] Single/Multi Prompt Agent Overview](https://docs.retellai.com/build/single-multi-prompt/prompt-overview) [[2] Retell AI’s Advanced Conversation Flow | Retell AI](https://www.retellai.com/blog/unlocking-complex-interactions-with-retell-ais-conversation-flow) [[3] Platform Changelogs | Retell AI](https://www.retellai.com/changelog) [[4] multi-prompt 2 parents for child state](https://discord.com/channels/1186424376300679209/1315288279234383963) [[5] Update Retell LLM](https://docs.retellai.com/api-references/update-retell-llm) ​

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

### Answer To improve transcription accuracy in Retell AI, you can use several approaches [(1) ]() : ## Use Boosted Keywords Add custom keywords to enhance the model's vocabulary for domain-specific terms . You can support up to 100 custom keywords . This is particularly helpful for special words, names, brands, or medical terminology that might be missing from transcripts . Configure boosted keywords in your agent settings [(2) ]() :

json
"boosted_keywords": ["retell", "kroger"]

## Adjust Denoising Mode If background noise or speech is affecting transcription quality, experiment with the denoising mode setting [(1) ]() . Retell offers different denoising options [(2) ]() :

json
"audio_processing": {
  "denoising_mode": "noise-cancellation"
}

Available denoising modes include

no-denoise

,

noise-cancellation

, and

noise-and-background-speech-cancellation

. For short responses like "sure" or "yes" being missed, if background noise isn't high, try setting denoising mode to "No Denoising" to preserve more of the raw audio signal [(1) ]() . ## Enable Transcription Mode for Accuracy Turn on transcription mode optimized for accuracy when sentences are being cut off prematurely . This helps prevent the transcription from finalizing sentences too early . ## Language Selection Ensure you've selected the correct language variant for your use case [(2) ]() . Retell supports 50+ languages [(3) ]() , including regional variants like

en-US

,

en-GB

, etc. [(2) ]()

json
"language": "en-US"

Retell's ASR has been significantly improved with lower Word Error Rates across multiple languages, particularly for European languages like German, French, Italian, and Polish, as well as Asian languages like Mandarin, Malay, and Hindi [(3) ]() .

​ To learn more, see the sources I considered: [[1] Increase transcription accuracy](https://docs.retellai.com/reliability/wrong-transcript) [[2] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) [[3] Smarter Understanding, Clearer Calls: Retell’s ASR Just Got a Major Upgrade | Retell AI](https://www.retellai.com/blog/retell-asr) ​

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