Understanding Filters

Filters help you generate the most relevant alerts. They allow you to limit alerts to only specific sets of objects, like entities (users or businesses), transactions, account numbers, addresses, countries, etc.

Filters are available in the Scenario Model Builder. A funnel icon will appear that allows agents granular control of the model logic:

Filters offer granular control over fields

The following list outlines some of the configurable filters:

Filter name	Description
`business`	Business entity information such as tax ID, name or registered location
`entity`	High level entity information such as status, tags, or type
`user`	User entity information such as ssn, gender, name, or DOB
`ip address`	IP address first or last seen by entity
`client fingerprint`	Fingerprint first or last seen by entity
`address`	Full address of an entity/event
`geolocation`	Last known IP address of an entity/event

Filters match for text and compare numbers

Put simply, filter expressions just check whether a value matches or doesn't match a different value.

In formal terms, these are called [boolean expressions (https://en.wikipedia.org/wiki/Boolean_expression).

Expression	Filter in practice
`business/name == Acme`	Only business named Acme
`business/name != Acme`	Only business NOT named Acme
`txn_amount > 10000`	Only transactions greater than ten thousand
`txn_amount < 10000`	Only transactions less than ten thousand
`entity/registered_at BETWEEN 01.01.2020 and 01.02.2020`	Only entities that registered between these days
`entity/address is empty`	Only entities without addresses
`entity/address is not empty`	Only entities with addresses
`last_name like %Smith%`	Only entities with names similar to Smith

📘
Use % when using the "Like" filter
When using Like or Not Like, be sure enclose % around the value. E.g, search for similarity with "ABC" using this %ABC% (see the following section for an example of a search)

Filters can be combined

You can also combine filters by using logical operators, like AND and NOT.

In this example we filters for entities:

who have a region:florida tag;
have a last name like Johnson AND have risk scores greater than 50 3) who DO NOT have empty SSNs.