Skip to main content
Back to Library
Prompt Engineering Guide

Mastering Customer support response
on Gemini 1.5 Pro

Stop guessing. See how professional prompt engineering transforms Gemini 1.5 Pro's output for specific technical tasks.

The "Vibe" Prompt

"Hey Gemini, can you help me write a customer support response for a user who's upset about a broken product they just received? Make it sound super empathetic and offer a solution. They're pretty mad, so be extra nice."
Low specificity, inconsistent output

Optimized Version

STABLE
Your task is to generate a customer support response. Below are the details about the customer's issue and guidelines for your response. ### Customer Issue Details: - **Customer Sentiment:** Extremely Upset - **Problem:** Received a newly purchased product (e.g., 'XYZ Widget') that is broken. - **Customer's Implied Need:** Resolution, apology, clear next steps. ### Response Guidelines: 1. **Acknowledge and Validate:** Start by acknowledging their frustration and validating their feelings. Use phrases that convey understanding. 2. **Apologize Sincerely:** Offer a direct and unambiguous apology for the inconvenience and the faulty product. 3. **State Empathy:** Explicitly express empathy for their situation. 4. **Propose a Solution:** Clearly outline the immediate next steps for resolution. This should typically involve a replacement or refund process. 5. **Reassure and Rebuild Trust:** Reassure them that this is not typical of your company's quality/service and that you are committed to making it right. 6. **Provide Call to Action:** Offer clear instructions on what they need to do or what will happen next (e.g., 'Please reply to this email with your order number,' 'We will ship a replacement within 2 business days'). 7. **Maintain Professional Tone:** Despite their anger, maintain a calm, professional, and courteous tone throughout. 8. **Avoid:** Jargon, making excuses, or blaming. **RESPONSE:**
Structured, task-focused, reduced hallucinations

Engineering Rationale

The optimized prompt leverages a chain-of-thought approach by breaking down the complex task of writing a support response into a series of logical steps and explicit instructions. It clearly defines the 'Customer Issue Details' to ensure the AI understands the context and sentiment. The 'Response Guidelines' act as a structured framework, guiding the AI to cover all necessary aspects of a good support interaction: acknowledgment, apology, empathy, solution, reassurance, and a call to action. It also explicitly lists what to avoid, pruning undesirable outputs. This structure reduces ambiguity and the need for the AI to infer requirements, leading to a more consistent, comprehensive, and high-quality output compared to the vague 'vibe_prompt.'

%
Token Efficiency Gain
The optimized prompt explicitly outlines the customer's sentiment and the product issue.
The optimized prompt provides a step-by-step guide for crafting the response, including acknowledgment, apology, solution, and reassurance.
The optimized prompt specifies the desired tone (professional, courteous) and what to avoid (jargon, excuses).

Ready to stop burning tokens?

Join 5,000+ developers using Prompt Optimizer to slash costs and boost LLM reliability.

Optimize My Prompts