> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.ntropy.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.ntropy.com/_mcp/server.

# Entities Identification

Entities identification is the first step in transaction enrichment. It identifies entities and classifies them according
to their **role** in the transaction and subsequent *type*. All entities returned as part of the enrichment are 
accessible via our [Entities API](/entities). 

# Roles
A transaction is composed of zero or more entities:
- An account holder that may or not be represented in the description.
- A counterparty that is usually present in the description, that can either be a `person` or an `organization`.
- One or more intermediaries. These are usually payment processing companies or food delivery services that act
on the behalf of the account holder and process their transactions.

# Account Holder Disambiguation
In some cases, we have transactions that may contain both the information about the account holder and the counterparty
in the same transaction. These ambiguities are notoriously difficult to resolve, specially in `business` transactions
where it's common for one or more organizations to appear in the description. 

Take the following transaction:
```
DEPOSIT CITY VENUE MGT | REF: LOAN | GA SILK AUD AUDLAUNDRY PTY LTD | *********** | *******************
```

From this description, we can extract `CITY VENUE MGT` and `SILK AUD AUDLAUNDRY PTY LTD` which are both organizations. 
Without more information, any could be the sender or the receiver and by extension the account holder
or the counterparty. Setting the `name` of the [account holder](./account-holders) (if you have it) could drastically
improve accuracy in cases like these.

---------------------------------------