Click Automations → Click New Automation → Click Answer all your FAQs.

The "Add your FAQs" section is the brain of your automation. When you click Add a FAQ, you are providing the raw data the AI uses to understand intent:

The Question
You don't need to type every variation of a question. Type the core intent (e.g., "Can I return a product?"). Zorcha’s AI is smart enough to recognize synonyms—so if someone asks "How do I send this back?" or "I don't like this, what's the refund policy?", it knows to use the answer you provided here.

The Answer

Input your standard, helpful response. You can include direct links so the user can take action immediately without needing further help.

Here are a few examples

This is where you calibrate the AI's personality to ensure every DM feels like an extension of your brand, not a generic computer response.

You aren't just giving the AI a name; you’re giving it a "job description".

You can name your persona → Social Media Assistant, Founder, Support Assistant).
​
This tells the AI how to identify itself. If you name it "Founder," the language will be more authoritative and personal. If you name it "Social Media Assistant," it will sound more helpful and supportive.
​

This determines the "vibe" of every word the AI writes.
​

Upbeat: High energy, enthusiastic, and very welcoming.

Neutral: Balanced, straightforward, and calm.

Formal: Polished, professional, and sophisticated.

Casual: Relatable, laid-back, and friendly.

"Upbeat" might say, "We’re so excited to help you!" while "Formal" would say, "We are pleased to assist you with your inquiry."

This ensures the AI respects your customer's time and your brand's style of talking.

Concise: Short, punchy, and straight to the point.

Balanced: Provides the answer with a bit of context.

Comprehensive: Detailed, thorough, and covers all possible follow-up points.
​

Emojis are the "body language" of digital text.

Emojis instantly make the AI feel more human and less like a sterile support ticket. They help emphasize points and show emotion that text alone cannot.
​

If a user asks a question that is outside the scope of your FAQs, the AI will send a "fallback" message that you can fully customize.

​Crucially, when this fallback is sent, the conversation remains marked as unread.

This ensures the customer isn't ignored, but you still get a notification to jump in and handle the query personally.

This automatically detects the language your user is using and generates the response in that exact same language. It removes the friction of translation. Whether your user DMs you in Spanish, French, or Hindi, they get a coherent answer immediately without you needing to hire a multilingual team.

You can toggle on a short, customizable note — often called a disclaimer — that appears at the end of every reply to indicate it was generated by AI.

This is highly recommended for businesses in sensitive industries like finance or healthcare, where users need to know they are interacting with an automated system.

It builds trust through transparency, letting your audience know that while the response is instant and helpful, it’s being powered by your digital assistant.

You’ll find a chat interface where you can talk to your AI exactly like a customer would. As you type questions, the AI will pull from the FAQ data and communication style you just defined—persona, tone, and all—to give you a real-time "sneak peek" of its performance.

The AI doesn't just look for exact keyword matches; it understands intent. When testing, you should ask questions in various ways to see how it adapts:

Original Question: "What is your return policy?"

Real World : "Hey, I don't really like this shirt, can I send it back?"

In both cases, the AI should recognize the intent as a "Return/Exchange" query and deliver the relevant answer you saved in "Add your FAQs".

To ensure your FAQs are robust, follow this three-pronged testing approach:

​Test the Accuracy
Ensure the links and specific details in your answers are correct.

​Test the Vibe
Check if the Persona and Tone feel right. Does the "Upbeat" assistant use too many emojis, or just enough?
​

Test the Boundaries
Ask a question that is not in your FAQs. Ensure the Fallback Message triggers correctly and that the chat stays marked as unread for your review.

Once you’ve chatted with your AI and it feels like a true extension of your brand, simply click Go Live. Your inbox is now officially on autopilot.