Skip to main content
Back to Library
Prompt Engineering Guide

Mastering Meeting notes extraction
on Claude 3.5 Haiku

Stop guessing. See how professional prompt engineering transforms Claude 3.5 Haiku's output for specific technical tasks.

The "Vibe" Prompt

"Extract the key decisions, action items, and participants from these meeting notes. Summarize the main discussion points. Meeting Notes: {meeting_notes}"
Low specificity, inconsistent output

Optimized Version

STABLE
You are an expert meeting assistant. Your task is to extract critical information from meeting notes. Follow these steps: 1. **Identify Meeting Metadata**: Extract the 'Meeting Title' (if present), 'Date' (if present), and list all unique 'Participants'. 2. **Summarize Main Discussion Points**: Condense the core topics and arguments discussed into 3-5 concise bullet points. 3. **Extract Key Decisions**: List all explicit decisions made during the meeting. For each decision, include the decision itself and any associated details (e.g., agreed-upon course of action, approval). 4. **Extract Action Items**: List all action items. For each action item, identify the 'Task', 'Assigned To' (the person responsible), and 'Due Date' (if specified). 5. **Output Format**: Present the extracted information in a structured JSON format with the following keys: `meeting_title` (string, optional), `date` (string, optional), `participants` (array of strings), `discussion_summary` (array of strings), `decisions` (array of strings), `action_items` (array of objects, each with `task`, `assigned_to`, `due_date`). If a field is not found, use an empty array or null. Meeting Notes: {meeting_notes}
Structured, task-focused, reduced hallucinations

Engineering Rationale

The optimized prompt provides clear, step-by-step instructions (chain-of-thought) to guide the model. It specifies the exact entities to extract and, crucially, defines a strict JSON output schema. This reduces ambiguity, ensures consistent output, and makes the extraction process more robust and easier for downstream systems to consume. The role definition ('expert meeting assistant') also helps set the context for the model's behavior. Explicitly stating optional fields and default values (empty array/null) for missing information further minimizes hallucination and improves reliability.

0%
Token Efficiency Gain
Optimized prompt consistently extracts decisions, action items, and participants.
Optimized prompt provides output in the specified JSON format.
Optimized prompt accurately identifies assigned individuals and due dates for action items.

Ready to stop burning tokens?

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

Optimize My Prompts