Loading Prompts...
Loading Prompts...
A guide to creating powerful system prompts for your AI agents at Gyld
System prompts are the instructions that define how your AI agent behaves. Think of them as the job description and training manual combined—they tell your AI agent who they are, what they should do, and how they should do it.
A well-crafted prompt can transform a generic AI into a specialized assistant that truly understands your business context and delivers consistent, high-quality results.
Key insight: The quality of your AI agent's output is directly tied to the quality of your prompt. Investing time in crafting a good prompt pays dividends in every interaction.
Vague instructions lead to unpredictable results. Instead of saying "be helpful," specify exactly what being helpful looks like in your context.
Give your AI agent a clear identity. What's their role? What's their communication style? Are they formal or casual? This consistency helps users know what to expect.
Clearly define what your AI agent should and shouldn't do. This prevents scope creep and ensures they stay focused on their designated tasks.
Provide relevant background information about your business, industry, or specific use case. The more context you provide, the more tailored the responses will be. With Gyld, you can also upload files to the Knowledge Base—documents, PDFs, and other resources that your AI agent can reference for deeper, more accurate context about your business.
Tell your AI agent how you want information presented. Should responses be concise or detailed? Bulleted or narrative? Professional or conversational?
A well-structured prompt typically includes these components:
1. Role Definition
Who is this AI agent? What's their job title and primary function?
2. Context & Background
Business details, industry specifics, team information
3. Task Instructions
What tasks should they perform? What's the workflow?
4. Output Requirements
Format, tone, length, and style expectations
5. Constraints & Boundaries
What they should avoid, limitations, escalation triggers
"You are an email assistant. Help with emails."
"You are Emma, a professional email assistant for a B2B SaaS company. Your role is to help draft, review, and organize emails. When drafting emails, maintain a professional but friendly tone. Keep responses concise (under 150 words for most emails). Always suggest a clear call-to-action. For customer-facing emails, emphasize helpfulness and solution-oriented language."
"Answer customer questions."
"You are Alex, a customer support specialist for [Company Name], a project management software. Your primary goal is to help users resolve issues quickly and leave them feeling valued. Start every interaction by acknowledging their concern. Provide step-by-step solutions when applicable. If you can't solve something, escalate to a human agent by saying: 'I want to make sure you get the best help possible. Let me connect you with our specialist team.' Never promise features or refunds without verification. Use the customer's name when possible."
"Be helpful" or "do a good job" don't give actionable guidance. Specify what "helpful" looks like in your context.
Asking for "detailed but brief" responses creates confusion. Be consistent in your expectations.
Without knowing your industry, company culture, or audience, the AI can't tailor responses appropriately.
Extremely long prompts with too many rules can be counterproductive. Focus on the most important guidelines.
Your first prompt won't be perfect. Test with real scenarios and refine based on the results.
Ready to create your AI agent? Start with a clear picture of what you need: