Token-Efficient Audits: Structured Comparison for Services, Travel, and Budgets

Comment utiliser cette collection

This collection provides a practical workflow for saving tokens by replacing open-ended queries with strict, table-based comparison prompts. Open-ended questions often lead to verbose, hallucinated responses that burn through your budget. By contrast, these prompts constrain the model to specific columns, forbid guessing, and require explicit labeling of unknowns. Step 1: Local Service Fairness Check. Use the 'Compare Local Service Listings' prompt when you have scattered text from websites or emails. Paste the raw listings. The prompt forces the AI to extract hours, fees, and eligibility into a fixed table. It will not average prices or invent phone numbers. This prevents the AI from writing a long, opinionated summary that wastes tokens on irrelevant fluff. You get a clean draft to review before making calls. Step 2: Travel Itinerary Cost Audit. Use the 'Travel Itinerary Time & Cost Comparison' prompt for complex trips. Instead of asking 'which trip is better?', paste the leg details (duration, cost, currency). The prompt calculates exact totals and differences in minutes and money. It flags unclear cost bases (e.g., per person vs. per vehicle) as 'unknowns' rather than guessing. This ensures you pay for precise arithmetic, not vague travel advice. Step 3: Electricity Tariff Analysis. Use the 'Electricity Tariff Comparison Analyst' prompt when reviewing new energy contracts. Paste your current tariff baseline (JSON) and the raw text from a new provider's quote. The prompt extracts fixed daily charges and unit rates, comparing them against your baseline. It explicitly flags ambiguous clauses like 'subject to market adjustment'. This prevents the AI from calculating projected bills based on assumed usage, ensuring you only pay for precise structural comparisons. Step 4: Household Budget Anomaly Detection. Use the 'Read-Only Budget Anomaly Brief' for monthly reviews. Paste transaction rows with headers. The prompt acts as a read-only auditor, flagging duplicates or outliers with confidence levels. It never edits files or executes payments. It produces a concise brief for human review, avoiding the high token cost of full financial analysis or open-ended chat about spending habits. Prerequisites: A plain-text AI chat interface. No API keys or software installation required. Limitations: These prompts are drafting and auditing tools. They do not browse live websites, check real-time prices, or execute bank transfers. Always verify final numbers with original source documents. Examples provided are fictional and for structural reference only.