← All solutions

EXPENSE CLASSIFICATION

Categorize expenses using your business’s own accounting labels.

Train AI on previously categorized transactions to suggest expense categories using your own accounting labels. Review results or integrate through an API.

FROM YOUR EXAMPLES

Monthly team collaboration software subscription
Software

Illustrative example · Not a live prediction

01

The same purchase can mean different things to different businesses

Consider this transaction: AWS — $847.22 A software company might classify it as Cloud Infrastructure. A consulting firm might use Software & Services. Another business may allocate it to a specific cost center. There is no single universal answer. The correct answer depends on how your business operates.

02

Your historical transactions can teach the model

If your accounting or expense system contains previous transactions and their assigned categories, you already have examples. For example:

Illustrative training examples
DescriptionCategory
AWS monthly invoiceCloud Infrastructure
Delta AirlinesBusiness Travel
Office DepotOffice Supplies
GitHubDeveloper Software

StayCharted AMT can train a model around those examples. Upload the historical data, identify the transaction description and the category you want predicted, and train.

03

Use AI where rules start getting complicated

Traditional automation often depends on rules such as: If description contains "AWS", select Cloud Infrastructure. That works until descriptions change. Or the same vendor appears in different situations. Or hundreds of rules accumulate over time. A trained model can use the patterns contained in previous examples rather than requiring someone to anticipate every future wording variation.

04

Keep people in control

Not every prediction has to be accepted automatically. For example:

  • High-confidence results can be applied automatically.
  • Medium-confidence results can be suggested to a user.
  • Low-confidence results can remain uncategorized for review.

That lets organizations choose the level of automation appropriate for their workflow.

05

Connect it to a finance application

For a small business, processing periodic exports may be enough. Accounting software providers, expense applications and internal finance systems can integrate the StayCharted AMT Prediction API directly.

  1. New Transaction
  2. Finance Application
  3. StayCharted AMT API
  4. Expense Category
  5. Suggest or Apply Category

StayCharted AMT supplies the prediction. Your application controls what happens with it.

STAYCHARTED AI MODEL TRAINER

Put your business knowledge to work.

Build your Business-Specific AI with AMT. Start with examples you already have, test your model’s results, and put it into your workflow.

Try AI Model Trainer