A transaction is a record of money that has been moved in or out of a bank account. Raw transaction information includes dates, amounts, and descriptions.

It's no easy task to extract extra information from these transactions! Inscribe, however, automatically extracts raw transaction data and adds additional context. This process, called transaction enrichment, derives additional information about transactions, specifically the following datapoints:

• Category
• Method
• Merchant

How are Transactions given Categories?

Raw transactions are parsed from each bank account and analyzed by our categorization engines to detect the relevant transaction category. A blend of rule-based and machine learning systems are used for transaction categorisation.

Inscribe has a defined transaction categorization taxonomy, which splits transactions into the following categories:

• Accommodation
  • Rent
  • Hotels
• Bank Charge
  • Overdraft
  • NSF
  • Interest Paid
• Entertainment
  • Gambling
• Loan
  • Credit Card Repayment
  • MCA (Merchant Cash Advance)
  • Mortgage
  • Payday
  • Auto Loan
  • Buy Now, Pay Later
• Income
  • Commission Received
  • Retirement
  • Revenue
  • Salary
  • Social Security
  • Interest Received
• Infrastructure
• Insurance
• Investments
  • Savings
  • Sweep Transfers
  • Trading
    • Crypto Trading
• Personnel
• Refund
• Retail
  • Food & Drink
  • Groceries
• Tax
• Transport
  • Fuel
• Utilities
• Withdrawal

When no transaction category can be detected, the transaction category is set to NULL.

How are Transactions given Methods?

A transaction method describes the way a transaction is processed. Similar to transaction categorization, a blend of rule-based and machine learning systems are used for transaction method categorisation. Inscribe extracts the following transaction methods:

• Bank Transfer
  • Self-Transfer
• Cash
• Check

When no transaction method can be detected, the transaction method is set to NULL.

How are Transactions given associated Vendors?

Many transactions have an associated vendor. For example, consider the following transaction description Purchased Flowers from Brett's Florist. In this case, Inscribe will extract Brett's Florist as the vendor. A combination of internal rules and named-entity recognition (NER) services are used to extract these vendors.

When no transaction vendor can be detected, the transaction vendor is set to NULL.