The instructions (also called the “prompt”) tell your Agent who it is, what it should do, and what it should avoid. A well-crafted prompt is the difference between an Agent that delivers consistent, high-quality results and one that produces unpredictable outputs.
The three components
Every Agent prompt consists of three sections:
1. Role / Purpose
Defines who the Agent is and what it’s trying to achieve. This sets the primary context for the LLM.
Keep this concise and focused on the end goal. One or two sentences is ideal.
You are a lead qualification specialist for Cargo. Your job is to evaluate
LinkedIn company URLs and determine if they match our Ideal Customer Profile.
2. Behavior
Outlines how the Agent should work—the steps it takes, tools it uses, and the format of its output. This directly guides the Agent’s planning phase.
Example:
1. Use the "Company-Lookup" tool to retrieve company data from the URL
2. Check the company against our ICP criteria in the Knowledge Base
3. Assign a qualification score: "High Fit", "Medium Fit", or "No Fit"
4. Provide a 2-3 sentence explanation of your reasoning
Tool selection tip: Give each tool a specific name and clear description.
The Agent uses these to decide which tool to call and when.
3. Avoidances
Establishes what the Agent must NOT do. These guardrails prevent undesired actions and ensure compliance.
Example:
- Never share proprietary ICP criteria with external parties
- Do not make up information if a lookup fails—report the failure instead
- Avoid overly casual language; maintain a professional tone
Avoidances are critical for security, compliance, and maintaining trust. Be
explicit about sensitive data handling and edge cases.
Best practices
| Practice | Why it matters |
|---|
| Be specific | Vague instructions lead to inconsistent results. Say “assign a score of 1-10” instead of “rate the lead” |
| Define output format | Specify exactly what the Agent should return (JSON, bullet points, a score, etc.) |
| List tools by name | Reference the exact tool names so the Agent knows what’s available |
| Handle edge cases | Tell the Agent what to do when data is missing or a tool fails |
| Test iteratively | Run test queries and refine the prompt based on actual outputs |
Example: Complete prompt
Here’s a full prompt for a lead qualification Agent:
ROLE:
You are a lead qualification specialist for Cargo. Evaluate inbound leads
and determine their fit based on our ICP criteria.
BEHAVIOR:
1. Extract the company domain from the lead's email address
2. Use "Enrichment-Tool" to gather company data (size, industry, funding)
3. Query the "ICP-Guidelines" resource to retrieve qualification criteria
4. Score the lead as "High", "Medium", or "Low" fit
5. Return a JSON object with: score, reasoning, and recommended next step
AVOIDANCES:
- Do not qualify leads from personal email domains (gmail, yahoo, etc.)
- Never fabricate company data—if enrichment fails, mark as "Needs Review"
- Do not recommend sales outreach for "Low" fit leads
