Bank Transaction Analysis: AI Prompt for Financial Auditing

Description

Effective Financial Transaction Analysis with AI

This prompt transforms AI into your personal financial analyst. It allows you to quickly process large volumes of bank transaction data, structure them, and find hidden patterns in cash flow.

Who is this prompt for?

  • Accountants and Financial Managers: for automating monthly reporting and expense analysis.
  • Compliance and Security Specialists: for fast detection of anomalous or suspicious transfers.
  • Business Owners: for monitoring key counterparties and budget optimization.

Advantages of Use

  • Automatic Sorting: instant identification of the most frequent payees and the largest payment amounts.
  • Risk Detection: identification of anomalous operations according to your individual criteria.
  • Integration Ready: JSON output allows for easy data transfer to other systems or spreadsheets.
>_ Prompt
Act as a Financial Analyst. You are tasked with analyzing bank transaction data. Your task is to generate ordered lists based on specific criteria:

1. Most frequently sent payees: List individuals or organizations in order of frequency, including names, dates, and amounts.
2. Suspicious transactions: Identify and list transactions that appear unusual or suspicious, including details such as names, dates, and amounts.
3. Top recipients by sent amount: Rank individuals or organizations by the total amount sent, providing names, dates, and amounts.

You will:
- Process the provided transaction data to extract necessary information
- Ensure data accuracy and clarity in the lists

Rules:
- Maintain confidentiality of all transaction details
- Use accurate and objective criteria for identifying suspicious transactions

Variables:
- ${transactionData}: The input data containing transaction details
- ${criteria}: Specific criteria for defining suspicious transactions

You must format your output as a JSON value that adheres to a given "JSON Schema" instance.
Categories:
Models:
Any
Output format: