### Answer Based on the available sources, it's not currently possible to directly convert a multi-prompt agent to a single-prompt agent (or vice versa) within the Retell AI platform. [(1) ]() You would need to manually create a new agent of the desired type and copy over your prompts and settings. The documentation does note that you should consider using a conversation flow agent or multi-prompt agent when your single prompt exceeds 1000 words or uses more than 5 functions. [(2) ]() If you're consolidating a multi-prompt agent into a single prompt, be aware of potential issues like hallucinations, unreliable function calling, maintenance challenges, and context confusion.

​ To learn more, see the sources I considered: [[1] Can you switch an Agent from a Single-prompt to a Multi-prompt (and vise versa)?](https://discord.com/channels/1186424376300679209/1325089959576866886) [[2] Single/Multi Prompt Agent Overview](https://docs.retellai.com/build/single-multi-prompt/prompt-overview) ​

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

### Answer --- Yes, I can help analyze prompts and provide best practice recommendations for Retell AI agents. [(1) ]() When sharing your prompt, I can review it against Retell AI's documented best practices, including: Structural recommendations: - Using sectional prompts with clear divisions like Identity, Style Guardrails, Response Guidelines, Task Instructions, and Objection Handling - Breaking large prompts into focused sections for better LLM comprehension and maintainability Tool calling guidance: - Defining explicit triggers and conditions for when functions should be called - Using exact function names and specifying sequences when tools should be called in order - Setting clear boundaries for when NOT to call certain tools General prompt engineering: - Ensuring tone and personality framing is clear [(2) ]() - Setting behavioral boundaries - Defining outcome prioritization - Including error handling guidance When to consider alternatives: - If your single prompt exceeds 1000 words or uses more than 5 functions, you should consider using a conversation flow agent or multi-prompt agent [(3) ]() - Complex logic with multiple decision branches may be better suited for Conversation Flow agents [(1) ]() Feel free to share your prompt and I'll provide specific recommendations based on these best practices.

​ To learn more, see the sources I considered: [[1] Prompt Engineering Guide](https://docs.retellai.com/build/prompt-engineering-guide) [[2] Prompt Engineering](https://www.retellai.com/glossary/prompt-engineering) [[3] Single/Multi Prompt Agent Overview](https://docs.retellai.com/build/single-multi-prompt/prompt-overview) ​

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

Thanks for the follow-up, @TheScrutiniser. I'll get back to you soon.

### Answer Your prompt has a solid structure and follows several Retell AI best practices. Here's my analysis with specific recommendations: ## Strengths **Sectional organization**: You've broken the prompt into clear sections (Identity, Style Guardrails, Response Guideline) , which aligns with Retell AI's recommended approach of using sectional prompts for better LLM comprehension and maintainability. [(1) ]() [(2) ]() **Concise tone guidance**: Your Style Guardrails effectively set expectations for conversational, concise responses—critical for voice agents. [(1) ]() **Specific behavioral boundaries**: You've included clear "don't" statements that help prevent unwanted phrases, which is good for maintaining brand voice. ## Recommendations for Improvement **Add explicit task steps**: Your Identity section has the conversation flow, but it would benefit from being reformatted as numbered steps with explicit "wait for user response" markers. This helps the LLM know when to stop and listen. [(2) ]() For example:

## Task
1. Introduce yourself: "Oh hi, it's Laura. I'm just calling from a car dealer near you about your car on Autotrader. Is it still for sale?"
wait for user response
2. If yes, ask: "Great, and are you open to offers before I get one of my team to call you back?"
wait for user response
3. If yes, ask: "Oh good, when's the best time?"
wait for user response
4. Collect the callback number and time.

**Clarify phone number pronunciation**: You mention repeating phone numbers "one-by-one" but could be more explicit about pronunciation. Consider adding guidance like the following to your Response Guideline section:

When speaking phone numbers, pronounce each digit individually with pauses:
- Transform formats like 07123456789
- Should be pronounced as: "zero seven one two three - four five six - seven eight nine"
- Don't omit the space around dashes when speaking

[(3) ]() **Date/time handling**: Your instruction about calculating "tomorrow" is good, but you could strengthen it by being more explicit about format. Consider adding:

For dates and times:
- For 3:00 PM, say "Three PM"
- For 8:45 AM, say "Eight forty-five AM"
- Always include "AM" or "PM"
- Use spoken date format: "Friday, January the fourteenth" not "1/14"

**Consider adding an Objection Handling section**: Since this is an outbound sales call, you might encounter common objections. Adding a section like this could help:

## Objection Handling
If the caller says they're not interested: "I understand. Just to confirm, you're no longer selling the car on Autotrader?"
If the caller is busy: "No worries at all. Would a different time work better for a quick chat?"

[(1) ]() **Tool/function calling**: If your agent uses any functions (like logging callback times, updating CRM, etc.) , you should add a "Tool Usage Instructions" section with explicit triggers for when to call those functions. ## Overall Assessment Your prompt is well-structured and concise. The main improvements would be: 1. Reformatting the conversation flow as explicit numbered steps 2. Adding more detailed pronunciation guidance for phone numbers and times 3. Including an objection handling section 4. Adding tool calling instructions if applicable The prompt should work well for a single-prompt agent, as it appears straightforward without complex branching logic that would require a conversation flow agent.

​ To learn more, see the sources I considered: [[1] Prompt Engineering Guide](https://docs.retellai.com/build/prompt-engineering-guide) [[2] Write a single prompt](https://docs.retellai.com/build/single-multi-prompt/write-single-prompt) [[3] Prompt guide & examples for specific situations](https://docs.retellai.com/build/prompt-situation-guide) ​

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