> ## Documentation Index
> Fetch the complete documentation index at: https://docs.famulor.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Prompt Guide for Specific Situations

> Specific prompt guides and examples for common scenarios when building a voice assistant

# Prompt Guide for Specific Situations

<Note>
  Here you will find specific prompt guides and examples for common scenarios you might encounter when building a voice assistant with Famulor.
</Note>

## Pronouncing Phone Numbers

<Card title="Use the Speak Slowly Feature" icon="volume-high" color="blue">
  We use the "speak slowly" feature to add pauses between words. For more information, see the <a href="/en/ai-assistants/example-prompts/pause-or-read-slowly">Speech Controllability Documentation</a>.
  It is recommended to use a prompt as a guideline for the LLM. This ensures the assistant consistently responds with the correct format, even when customers verify the phone number.
</Card>

### Example for Pronouncing Phone Numbers

<CodeGroup>
  <CodeGroupItem title="Phone Number Prompt">
    ```markdown theme={null}
    When people ask for your phone number, your phone number is 03012345678

    ## Guidelines
    When pronouncing the phone number, transform the format as follows:
    - Input formats like 03012345678, (030) 123-45678, or 030-123-45678
    - Should be pronounced as: "null drei null - eins zwei drei - vier fünf sechs sieben acht"
    - Important: Do not omit the space around the dash when speaking
    ```
  </CodeGroupItem>

  <CodeGroupItem title="German Pronunciation">
    ```markdown theme={null}
    ## German Phone Numbers

    For German phone numbers:
    - 0 is pronounced as "null"
    - Use dashes for better clarity
    - Example: 030-123-45678 → "null drei null - eins zwei drei - vier fünf sechs sieben acht"
    ```
  </CodeGroupItem>
</CodeGroup>

## Spelling Email Addresses

<Card title="Pronounce Email Addresses Correctly" icon="envelope" color="green">
  Email addresses must be spelled out to avoid confusion. Here is the correct approach:
</Card>

### Example for Spelling Email Addresses

<CodeGroup>
  <CodeGroupItem title="Email Format">
    ```markdown theme={null}
    ## How to spell

    The possible email format is name@unternehmen.de

    To spell an email address:
    n-a-m-e-@-u-n-t-e-r-n-e-h-m-e-n-dot-de

    @ is pronounced as "at"
    ```
  </CodeGroupItem>

  <CodeGroupItem title="German Email Addresses">
    ```markdown theme={null}
    ## German Email Addresses

    For German email addresses:
    - Use German pronunciation for letters
    - Example: info@famulor.de → "i-n-f-o-at-f-a-m-u-l-o-r-punkt-d-e"
    - Dot is pronounced as "punkt"
    ```
  </CodeGroupItem>
</CodeGroup>

## Pronouncing Websites

<Card title="Pronounce Website URLs Correctly" icon="globe" color="purple">
  Website URLs must be pronounced clearly and understandably. Here is the correct approach:
</Card>

### Example for Pronouncing Websites

<CodeGroup>
  <CodeGroupItem title="General Guidelines">
    ```markdown theme={null}
    Whenever you encounter a website URL, please:

    1. Identify each segment of the domain name
    2. If a segment consists of individual letters (e.g., "NK"), pronounce each letter in its spoken form in German (e.g., "N" → "en", "K" → "ka")
    3. If a segment is a recognizable word (e.g., "laundry"), pronounce it normally as that word
    4. Pronounce "punkt" before the top-level domain (e.g., "punkt de", "punkt net", "punkt org", etc.)
    ```
  </CodeGroupItem>

  <CodeGroupItem title="German Examples">
    ```markdown theme={null}
    ## German Examples

    "famulor.de" → "f-a-m-u-l-o-r punkt de"
    "testfirma.net" → "t-e-s-t-f-i-r-m-a punkt net"
    "beispiel.org" → "b-e-i-s-p-i-e-l punkt org"

    Stick exactly to this phonetic breakdown to ensure clarity and correct pronunciation for customers.
    ```
  </CodeGroupItem>
</CodeGroup>

## Pronouncing Times

<Card title="Pronounce Times Correctly" icon="clock" color="orange">
  Times must be pronounced in an understandable format. Here are the guidelines:
</Card>

### Example for Pronouncing Times

<CodeGroup>
  <CodeGroupItem title="English Times">
    ```markdown theme={null}
    ## For times, numbers, and dates

    For 1:00 PM, say "One PM"
    For 3:30 PM, say "Three thirty PM"
    For 8:45 AM, say "Eight forty-five AM"

    Never say "O'clock"; instead, say "O-Clock"
    Always say "AM" or "PM"
    ```
  </CodeGroupItem>

  <CodeGroupItem title="German Times">
    ```markdown theme={null}
    ## For times

    For 13:00, say "Dreizehn Uhr" or "Ein Uhr nachmittags"
    For 15:30, say "Fünfzehn Uhr dreißig" or "Drei Uhr dreißig nachmittags"
    For 8:45, say "Acht Uhr fünfundvierzig" or "Viertel vor neun morgens"

    Use "morgens" for AM and "nachmittags" for PM
    ```
  </CodeGroupItem>

  <CodeGroupItem title="Alternative Formats">
    ```markdown theme={null}
    ## Alternative formats

    You can also say:
    - "Halb neun" for 8:30
    - "Viertel nach zwei" for 14:15
    - "Drei viertel vier" for 15:45

    Choose the format that is most understandable for your customers.
    ```
  </CodeGroupItem>
</CodeGroup>

## Putting Customers On Hold / No Response Needed

<Card title="Putting Customers on Hold" icon="pause" color="yellow">
  Sometimes you need to put customers on hold or no response is required. Famulor has a special function for this:
</Card>

### For Non-Reasoning Models

<CodeGroup>
  <CodeGroupItem title="Use NO_RESPONSE_NEEDED">
    ```markdown theme={null}
    ## Putting on hold

    Famulor has a hardcoded stop sequence in the LLM: `NO_RESPONSE_NEEDED`.
    Whenever this sequence is detected, response generation stops.

    You can instruct the LLM to output nothing by writing:

    - When the user says "one moment" or "please hold", respond exactly with: "NO_RESPONSE_NEEDED"
    ```
  </CodeGroupItem>

  <CodeGroupItem title="Example">
    ```markdown theme={null}
    ## Example Implementation

    When the customer says:
    - "one moment"
    - "wait a second"
    - "please hold"

    Then respond exactly: "NO_RESPONSE_NEEDED"

    Response generation stops and the assistant waits for further input.
    ```
  </CodeGroupItem>
</CodeGroup>

### For Reasoning Models

<Warning>
  Reasoning models like `gpt-5` or `gpt-5.1` do not support this feature. You must craft prompts differently to achieve this behavior.
</Warning>

<CodeGroup>
  <CodeGroupItem title="Alternative for Reasoning Models">
    ```markdown theme={null}
    ## Putting on hold for Reasoning Models

    When the user says "one moment" or similar, simply do not respond.
    Do not react to hold signals, just wait for further instructions.
    ```
  </CodeGroupItem>

  <CodeGroupItem title="Example Prompt">
    ```markdown theme={null}
    ## Example Prompt

    When the user says:
    - "one moment"
    - "wait a second"
    - "please hold"

    Then DO NOT respond. Just wait until the user speaks again.
    ```
  </CodeGroupItem>
</CodeGroup>

## Normalizing Text for Speech

<Card title="Normalize Text for Speech" icon="font" color="indigo">
  Normalize certain parts of the text (numbers, currency, dates, etc.) into their spoken form for more consistent speech synthesis. Sometimes TTS models misread non-normalized text.
</Card>

### Example for Text Normalization

<CodeGroup>
  <CodeGroupItem title="Before Normalization">
    ```markdown theme={null}
    ## Before normalization

    Call my number 03012345678 on July 5, 2024 for the payment of €24.12
    ```
  </CodeGroupItem>

  <CodeGroupItem title="After Normalization">
    ```markdown theme={null}
    ## After normalization

    Call my number null drei null eins zwei drei vier fünf sechs sieben acht on the fifth of July twenty twenty-four for the twenty-four Euro twelve Cent payment
    ```
  </CodeGroupItem>
</CodeGroup>

<Note>
  This feature adds a small latency (\~100ms) to the overall process.
</Note>

## Language Support

<Card title="Supported Languages" icon="globe" color="teal">
  Currently, text normalization is supported for the following languages:

  * **German** - full support
  * **English** - full support
  * **Spanish** - full support
  * **French** - full support

  Other languages result in a no-op (text is not modified) for this step.
</Card>

<Card title="Language Settings" icon="gear" color="blue">
  When you select a language other than multilingual, that language code is used to normalize the text (e.g., 1 is normalized to "eins" when using German).
  When you select multilingual, the language is detected automatically based on the generated text and normalized accordingly.
</Card>

## Prompt Engineering Guide

<Card title="Best Practices for Effective Prompts" icon="book-open" color="blue">
  Best practices for writing effective prompts that create reliable and natural-sounding AI phone assistants.
</Card>

### Introduction

Prompt engineering is the foundation for creating effective AI phone assistants. A well-designed prompt determines how your assistant interprets situations, responds to users, and handles edge cases. This guide provides proven strategies for writing prompts your assistant can reliably follow.

<Note>
  This guide focuses on general prompt engineering principles. For agent-specific implementation:

  * **Single/Multi Prompt Agents**: Apply these principles directly in your prompts
  * **Conversation Flow Agents**: Use these principles within individual node instructions
</Note>

### Best Practice 1: Use Structured Prompts

Break large prompts into focused sections to improve organization and LLM understanding. This structured approach offers several advantages:

* **Reusability**: Sections can be adapted across different agents
* **Maintainability**: Easily update specific behaviors without affecting others
* **Clarity**: LLMs process structured information more accurately

#### Recommended Prompt Structure

```markdown theme={null}
## Identity

You are a friendly AI assistant for [Company Name].

Your role is [specific purpose].

You have expertise in [relevant areas].

## Style Guidelines

Be concise: Keep answers under 2 sentences unless explaining complex topics.

Be conversational: Use natural language, contractions, and acknowledge what the caller says.

Be empathetic: Show understanding for the caller's situation.

## Response Guidelines

Return data in spoken form: say "fifteenth January" not "15.01".

Ask one question at a time: avoid overwhelming the caller with multiple questions.

Confirm understanding: paraphrase key information back to the caller.

## Task Instructions

[Specific steps the assistant should follow]

## Objection Handling

If the caller says they are not interested: "I understand. Is there something specific..."

If the caller is frustrated: "I understand your frustration, let me help solve this..."
```

### Best Practice 2: Use Conversation Flow for Complex Tasks

If your assistant needs to handle complex logic or multiple tools, use Conversation Flow Agents instead of managing everything in a single prompt.

#### When to switch to Conversation Flow:

* **Multiple decision branches**: More than 3-4 conditional paths
* **Tool coordination**: Using 5+ different functions/tools
* **State management**: Tracking multiple variables over conversation
* **Reliability concerns**: Single prompt exhibits inconsistent behavior

#### Benefits of Conversation Flow:

* Each node focuses on one specific task
* Deterministic tool call and transitions
* Easier to debug and optimize individual steps
* More predictable assistant behavior

### Best Practice 3: Explicit Tool Call Instructions

<Note>
  This section applies only to Single/Multi Prompt Agents. Conversation Flow Agents handle function calls deterministically via their node configuration.
</Note>

#### The Challenge

LLMs often struggle to determine when to call tools based only on tool descriptions. Without explicit instructions, agents may:

* Call tools at inappropriate times
* Fail to call tools when needed
* Use the wrong tool for a situation

#### Solution: Define Clear Triggers

Always provide precise conditions for tool usage in your prompts. Refer to tools by their exact names.

##### Example: Customer Service Agent

```markdown theme={null}
## Tool Usage Instructions

1. First, gather information about the customer's issue.

2. Determine the type of request:

   - If the customer mentions "refund" or "money back":

     → Call the function `transfer_to_support` immediately

   - If the customer needs order status:

     → Call the function `check_order_status` with order_id

   - If the customer wants to change their order:

     → First call `check_order_status`

     → Then switch to change status

3. After retrieving information:

   - Always summarize what you found

   - Ask if they need additional help

   - If yes, determine the next appropriate action
```

#### Best Practices for Tool Instructions

1. **Use trigger words**: List specific words/phrases that should trigger tool calls
2. **Define sequences**: Specify when tools should be called in order
3. **Set boundaries**: Clarify when certain tools SHOULD NOT be called
4. **Provide context**: Explain why each tool is called

## Practical Examples for Famulor

<Card title="Famulor-Specific Adjustments" icon="star" color="red">
  Here are some Famulor-specific customizations:
</Card>

### Example of a Complete Prompt

<CodeGroup>
  <CodeGroupItem title="Complete Example Prompt">
    ```markdown theme={null}
    # Famulor Assistant - Prompt for Specific Situations

    ## Identity
    You are a professional Famulor AI assistant supporting customers on phone calls.

    ## Pronunciation Guidelines

    ### Phone Numbers
    - Use German pronunciation
    - Example: 030-123-45678 → "null drei null - eins zwei drei - vier fünf sechs sieben acht"

    ### Email Addresses
    - Spell every part out
    - Example: info@famulor.de → "i-n-f-o-at-f-a-m-u-l-o-r-punkt-d-e"

    ### Websites
    - Pronounce each letter individually
    - Example: Famulor.io → "f-a-m-u-l-o-r-punkt-d-e"

    ### Times
    - Use German time formats
    - Example: 14:30 → "Vierzehn Uhr dreißig"

    ## Behavior
    - Be polite and professional
    - Use the pause feature when needed
    - Normalize numbers and dates for better clarity
    ```
  </CodeGroupItem>
</CodeGroup>

## Testing and Optimizing

<Card title="Test Your Prompts" icon="play" color="green">
  * **Conduct audio tests** - Listen to the pronunciation
  * **Test various scenarios** - Cover all specific situations
  * **Collect customer feedback** - Ask about understandability
  * **Iteratively improve** - Adjust based on results
</Card>

## Help and Support

<Tip>
  Our support team can help you customize your prompts for specific situations.
</Tip>

<Card title="Contact Support" icon="headphones" color="blue" href="mailto:support@famulor.io">
  Need help customizing your prompts? Contact our support team.
</Card>

## Next Steps

<Card title="Further Resources" icon="arrow-right" color="purple">
  * Check out our <a href="/en/ai-assistants/example-prompts/general-prompt-engineering-guide">General Prompt Engineering Guide</a>
  * Explore our **Prompt Templates** in the dashboard
  * Learn more about <a href="/en/ai-assistants/example-prompts/pause-or-read-slowly">Speech Controllability</a>
  * Test your prompts with **Audio Tests**
</Card>

<Tip>
  Related pages: [What Is AI Assistant](/en/ai-assistants/what-is-ai-assistant) and [Prompt And Tools](/en/ai-assistants/prompt-and-tools), and [Assistant Best Practices](/en/ai-assistants/assistant-best-practices).
</Tip>