Structured Cost Audits: Electricity, Travel, Returns & Budgets

Comment utiliser cette collection

This collection provides a practical workflow for saving tokens and preventing costly errors 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: Electricity Tariff Audit. Use the 'Electricity Tariff Comparison Analyst' prompt when switching providers. Paste your current tariff JSON baseline and the raw text from a new contract. The prompt extracts fixed daily charges and unit rates (including peak/off-peak if applicable). It compares these against your baseline and flags ambiguous clauses like 'subject to market adjustment'. It does not calculate projected bills based on assumed usage, ensuring you only pay for precise structural comparisons. Step 3: E-Commerce Return Policy Check. Use the 'E-Commerce Return Policy Auditor' before initiating a return. Paste the receipt text or order confirmation email. The prompt extracts critical fields: return window, start date trigger, condition requirements, restocking fees, and return shipping costs. It calculates estimated deadlines and flags risks like 'buyer pays shipping' or '10% restocking fee'. It never contacts sellers or generates labels; it provides a risk assessment to help you decide if the return is financially viable. 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 transactions. Always verify final numbers with original source documents. Examples provided are fictional and for structural reference only.