The AI uses only what you write in the policy field. It will not invent refunds, discounts, timelines or support links you didn't provide — keeping your replies accurate and safe to send.
Situation-aware empathy
The AI adjusts the opening based on the customer situation you select — validating feelings for frustrated customers, being factual for technical queries, casual for general questions.
No robotic language
The prompt bans clichés like "We apologize for the inconvenience" and "Thank you for reaching out", forcing replies that sound like a real person wrote them.
Channel-appropriate length
The prompt adapts length to the channel — concise for live chat and social comments, fuller for email and help desk tickets — reducing back-and-forth and closing tickets faster.
Tip: If you're denying a refund or a feature request, choose "Firm, polite, and policy-driven" tone. The AI will deliver the no clearly without leaving room to argue — while still sounding human.