Skip to main content

Response Settings

Response Settings is where you teach the AI how to write replies that sound like you. This is the heart of Rewo—these settings control the AI's personality, knowledge, and behavior to ensure every response matches your brand voice and provides accurate information.

The settings are organized into four tabs: Response Behavior, Knowledge & Context, Personality, and Advanced Settings. These tabs are arranged from the most basic to the most advanced features, allowing you to configure the AI step by step.

warning

Changes to these settings should be tested carefully to avoid undesired outcomes. Adjusting advanced settings without understanding their impact may lead to unexpected behavior in the AI's responses.

tip

To get a feeling for how your changes to the settings affect responses consider temporarily pausing auto drafting responses and generating a few draft responses via the dashboard yourself. Simply re-generate responses for a review to see if the changes are what you expect and if AI behavior is consistent. Careful not to accidentally save or publish a response in that process.

Localization

Responses are automatically generated in the language of the reviewer. As such you may need to use the translation toggle to view a response in English.

Response Behavior​

The Response Behavior tab allows you to control how the AI structures and formats its responses. These settings help ensure that the tone and style of the replies align with your brand's communication preferences.

Include Greeting in Responses​

This setting determines whether responses start with a greeting, such as "Hi [name]," or "Hello,". The greeting style (whether formal or casual) is automatically determined by your AI personality settings.

Localization

Don't worry about your user's languages here. Your greeting will be automatically translated into an appropriate localized version.

Signature Settings​

The Signature Settings control when and how a signature is included at the end of responses. You can choose from the following options:

  • AI decides when appropriate (auto): The AI includes a signature when it adds value, such as in formal or important responses, while skipping it for brief, casual replies.
  • All responses (all): Every response ends with your signature, ensuring consistency but potentially feeling repetitive for short replies.
  • Never (none): No signatures are added, which is best for brands that prefer brief, informal communication.

The signature text itself should be concise and brand-appropriate. Examples include:

  • "The [App Name] Team"
  • "Best regards, Sarah from Support"
  • "Cheers! — The [Company] Crew"
Localization

Don't worry about your user's languages here. Your signature will be automatically translated into an appropriate localized version.

Support Email Configuration​

This setting determines when to include your support email address in responses. You can configure it as follows:

  • AI decides when appropriate (auto): The AI includes your email when additional help is likely needed or when it asks users to reach out or provide more details on a complaint.
  • All responses (all): The support email is included in every response, regardless whether it is actually needed or not.
  • Negative reviews only (negative): The email is included only for low-rated reviews, where users are more likely to need direct assistance.
  • Never (none): The support email is never included, which is suitable if you prefer users to contact you through other channels.

Knowledge & Context​

The Knowledge & Context tab is where you provide the AI with detailed information about your app and your brand. The more comprehensive and accurate the information you provide, the better the AI will be at generating helpful and relevant responses.

Brand & Communication​

The Brand Voice setting allows you to define your brand's overall communication style and personality. This includes details such as your tone (e.g., "friendly and approachable" or "professional and authoritative"), key values, and any specific communication guidelines. For example:

"We're a playful, user-first brand that values transparency and simplicity. We speak directly and honestly with users, avoiding corporate jargon. We're enthusiastic about our product but never pushy. Think: helpful friend, not salesperson."

General Context​

The General Context setting provides broader information about your company. This helps the AI understand your business and tailor responses accordingly. Include details such as what your app does, your target audience, and your unique selling points. For example:

"We're a fitness tracking app designed for busy professionals who want simple, effective workouts. Founded in 2020, we focus on 15-30 minute home workouts that require no equipment. Our user base is 70% women aged 25-45. We're a small team of 8 passionate about making fitness accessible to everyone."

Product Information​

The Product Information section allows you to provide detailed knowledge about your app's features, known issues, limitations, and roadmap items. This ensures the AI can address user questions and concerns accurately.

Product Features​

List your app's key features, including a brief description of what each feature does and the benefits it provides. For example:

  • "Offline Mode: Use workouts without internet; syncs when online."
  • "Custom Playlists: Share workout playlists with Spotify or Apple Music."
  • "Progress Photos: Securely store photos to track fitness progress."

Known Issues​

Document any current bugs or problems you're aware of and working on. This helps the AI acknowledge these issues and provide updates or workarounds. For example:

  • "Sync delays with Apple Health on iOS 17.2. We're working on a fix. Workaround: manually refresh in Settings."
  • "Some Android users see slow loading on workout history. Fix scheduled for v2.3 release in early Feb."

Known Limitations​

List features or capabilities your app doesn't currently have. This helps set appropriate expectations for users. For example:

  • "No Apple Watch standalone app currently. You can mirror workouts from your iPhone to your Watch."
  • "We don't support Garmin integration yet. It's on our roadmap for Q2 2025."

Roadmap Items​

Share upcoming features or improvements you're planning. This helps the AI inform users about what's coming soon. For example:

  • "Social features launching Q1 2025: follow friends, share workouts, compete on leaderboards."
  • "AI workout generator coming soon: personalized workouts based on your goals, time, and equipment."

Personality​

Fine-tune the AI's personality traits to match your brand voice. Each slider ranges from -2 to +2, with 0 as neutral or default. These settings shape how the AI writes, from word choice to sentence structure.

warning

Careful with changes here as they can be very impactful on your generated responses. Selecting contradicting or inconsistent values across multiple personality traits may lead to unexpected AI behavior.

  • Formality: Formality defines the tone and style of the AI's language. It determines whether the AI uses professional, polished language or a more casual, conversational tone. Adjust this setting based on your audience's expectations and your brand's voice.
  • Conciseness: Conciseness controls the level of detail in responses. You can choose whether the AI provides brief, direct answers or more detailed, comprehensive explanations. This is useful for tailoring responses to your users' needs, such as mobile users preferring brevity or complex products requiring detail.
  • Empathy: Empathy shapes how the AI acknowledges and responds to user emotions. It can range from being straightforward and fact-focused to showing deep understanding and validation of feelings. This is particularly important for user-facing apps where emotional sensitivity is key.
  • Enthusiasm: Enthusiasm adjusts the energy and positivity in the AI's tone. It can range from calm and measured to excited and energetic. Choose a level that aligns with your brand's personality and the context of your communication.
  • Helpfulness: Helpfulness determines how much additional information or proactive suggestions the AI includes in its responses. This is ideal for educational or complex apps that benefit from extra guidance, while simpler apps may prefer a more direct approach.
  • Conciliation: Conciliation defines how apologetic or conflict-resolving the AI is. It can range from stating facts without unnecessary apologies to actively smoothing over conflicts and prioritizing harmony. This is useful for customer service-focused apps.
  • Humor: Humor adds a playful and engaging tone to responses. It should be used cautiously, as it may not suit all contexts or audiences. Fun, casual brands can benefit from humor, while professional or sensitive topics should avoid it.
  • Credibility: Credibility ensures the AI's tone matches the level of professionalism required. It can range from friendly and approachable to expert and authoritative. This is crucial for industries like finance or healthcare, where trust is paramount.
  • Authenticity: Authenticity makes the AI sound natural and genuine, avoiding overly scripted or corporate language. Most brands benefit from a higher level of authenticity, as users appreciate responses that feel human and relatable.
  • Emoji Usage: Emoji usage determines how frequently emojis appear in responses. This can range from no emojis to regular, tasteful use. Younger audiences and casual brands may appreciate more emojis, while professional contexts may require fewer or none.

Advanced Settings​

The Advanced Settings tab provides fine-grained control for power users who want to tune the AI beyond the standard personality and behavior rules. These settings allow you to influence the AI's learning patterns, enforce specific terminology, and even provide direct system instructions for complex scenarios.

Experience Influence​

This setting balances your rule sets and configuration against your historical published responses. By analyzing your history, the AI can catch nuances and patterns specific to your brand that manual settings might miss.

  • Fresh approach (0): The AI relies strictly on your current rules and ignores past history. Every response starts fresh based on your settings.
  • Learn from history (1): The AI heavily references your past replies, mimicking the specific style, tone, and patterns found in your published history.
warning

High weights allow historical patterns to take precedence over any other setting, including your Brand Voice, Personality, Phrases, and Custom Instructions. If your history contains repetitive patterns or undesired, the AI will replicate them regardless of your rules. Ensure your published history reflects the quality you want to see.

tip

If the AI starts sounding too much like your old responses and isn't adapting to your new settings, lower this value.

Phrases​

The Phrases section allows you to enforce specific terminology or exclude language that doesn't align with your brand. This ensures consistency across all AI-generated responses while avoiding overused clichés or insensitive terms.

Golden Phrases​

List keywords or sentences the AI should try to include in responses when contextually appropriate. These will only be used if they fit naturally into the flow of the conversation. For example:

  • "We're here to help"
  • "We're glad to hear you're enjoying the new update."
  • "We appreciate you sharing your experience with us."

Banned Phrases​

Specify terms or phrases the AI should actively avoid in any response. The AI will find alternative ways to convey the same meaning without using these exact sequences. For example:

  • "It's not our fault"
  • "As per our policy,"
  • "Your feedback has been noted."

Custom Prompt​

The Custom Prompt feature allows you to provide direct instructions to the AI system, giving you fine-grained control over its behavior beyond the standard rules and sliders.

Position​

This setting determines where your instructions are placed within the final AI prompt. This influences how much weight your instructions carry during response generation.

  • Before system prompt: Your instructions are injected before the examples and the actual review details. This guides the AI while allowing it to integrate everything naturally.
  • After system prompt: Your instructions are added at the very end of the prompt, ensuring they are the last thing the AI considers before writing.
  • Replace system prompt: Your instructions completely replace Rewo's personality and behavior logic. In this mode, all Brand Voice, Personality sliders, and Response Behavior settings are bypassed. However, the Knowledge Base (product info, known issues, etc.), examples, and the actual review details remain available to the AI.
info

When using Replace, only Rewo's core system logic remains active to ensure the AI understands its primary task of responding to a review. You take full responsibility for defining the tone and rules of the response.

Custom Instructions​

Provide your actual instructions to the AI (up to 2,000 characters). This is useful for enforcing very specific response patterns or handling unique edge cases. For example:

"ALWAYS acknowledge the specific feature the user mentioned. If they report a bug, thank them and explain it helps us improve. NEVER promise exact timelines for fixes. IF users mention competitors, stay positive, NEVER criticize other apps."