How rule-based Identity Resolution works in Amazon Connect

This topic describes how rule-based Identity Resolution performs automatic profile matching and how it automatically merges similar profiles.

Automatic profile matching

To identify similar profiles, the rule-based Identity Resolution uses a list of matching rule attributes to match each profile. Up to 15 MatchingRule attributes are supported in the MatchingRules.

Matching rules

The following is a list of the MatchingRule attributes that can be used. You can configure up to 15 matching rule levels. For each matching rule, you can use the following Personal Identifiable Information (PII) attributes in each profile:

• AccountNumber

• Address.Address: All addresses specified in the Attribute Type Selector are reviewed for similarity, including Address, BusinessAddress, MaillingAddress, and ShippingAddress

• Address.City: All addresses specified in the Attribute Type Selector are reviewed for similarity, including Address, BusinessAddress, MaillingAddress, and ShippingAddress

• Address.Country: All addresses specified in the Attribute Type Selector are reviewed for similarity, including Address, BusinessAddress, MaillingAddress, and ShippingAddress

• Address.County: All addresses specified in the Attribute Type Selector are reviewed for similarity, including Address, BusinessAddress, MaillingAddress, and ShippingAddress

• Address.PostalCode: All addresses specified in the Attribute Type Selector are reviewed for similarity, including Address, BusinessAddress, MaillingAddress, and ShippingAddress

• Address.State: All addresses specified in the Attribute Type Selector are reviewed for similarity, including Address, BusinessAddress, MaillingAddress, and ShippingAddress

• Address.Province: All addresses specified in the Attribute Type Selector are reviewed for similarity, including Address, BusinessAddress, MaillingAddress, and ShippingAddress

• PhoneNumber: The phone numbers specified in the Attribute Type Selector are reviewed for similarity, including PhoneNumber, HomePhoneNumber, and MobilePhoneNumber.

• EmailAddress: All email addresses specified in the Attribute Type Selector are reviewed for similarity, including EmailAddress, BusinessEmailAddress, and PersonalEmailAddress

• BirthDate

• BusinessName

• FirstName

• LastName

• MiddleName

• Gender

• Any customized profile attributes with the Attributes prefix

Matching rules are processed by priority. For example, the first rule should be the most optimized rule you want to define and should be used to achieve the most accurate result.

Attribute Type Selector

The Attribute Types Selector holds vital configuration information for rule-based identity resolution, facilitating profile matching. This allows you to fine-tune the comparison of profiles across attribute types and select the key attributes for matching within each type. Within this feature, you have the flexibility to configure three distinct attribute types, enabling precise control over the matching process.

• Email type

  • You can choose from EmailAddress, BusinessEmailAddress, and PersonalEmailAddress

• PhoneNumber type

  • You can choose from PhoneNumberNumber, HomePhoneNumber, and MobilePhoneNumber

• Address type

  • You can choose from Address, BusinessAddress, MaillingAddress, and ShippingAddress

You can either choose ONE_TO_ONE or MANY_TO_MANY as the AttributeMatchingModel. When choosing MANY_TO_MANY, the system can match attributes across the sub-types of an attribute type. For example, if the value of the EmailAddress field of Profile A and the value of BusinessEmailAddress field of Profile B matches, the two profiles are matched on the EmailAddress type. When choosing ONE_TO_ONE, the system can only match if the sub-types are exact matches. For example, only when the value of the EmailAddress field of Profile A and the value of the EmailAddress field of Profile B matches, the two profiles are matched on the EmailAddress type.

Max allowed rule level for matching

You can configure the max rule level you want to use to match similar profiles. For example, if your max allowed rule level for matching is 5, the system will not find similar profiles using the rule level 6.

Match groups

A match group consists of all similar profiles which represent a customer. Each match group contains the following information:

• A match ID, which uniquely identifies the group of two or more similar profiles that represent a contact

• The number of profile IDs in the match group

Match status

• PENDING

  The first status after the configuration of a rule-based Matching rule. If it is an existing domain, the rule-based Identity Resolution waits 1 hour before creating the matching rule. If it is a new domain, the system will skip the PENDING stage.

• IN_PROGRESS

  The system is creating the rule-based Matching rule. Under this status, the system is evaluating the existing data and you can no longer change the rule-based matching configuration.

• ACTIVE

  The rule is ready to use. You can change the rule a day after the status is in an ACTIVE state.

How the auto-matching process works

After you create a new Amazon Connect Customer Profiles domain with the rule-based matching rule, the rule-based Identity Resolution will match similar profiles based on the rule you specified while you are ingesting the profiles. If you update the configuration of the rule-based matching, Customer Profiles will start to re-evaluate the profiles in your domain using the new configuration in an hour.

If you are enabling the rule-based matching with an existing domain, the system will move to a PENDING state and will start to evaluate the existing profiles in your domain using the new configuration in an hour. The time it takes to finish evaluating profiles is dependant on how many profiles exist.

• By default, the default rule will be applied if no custom rule specified.

  • Amazon Connect Customer Profiles provides a default matching rule if you do not provide a custom matching rule. You can check the custom matching rule here.

• All the records will go through the rule-based matching rules.

  • The system assesses each matching rule level until a match is identified or until the maximum allowed rule level for matching is reached. The evaluation process commences at rule level 1, where the record is analyzed. If no matching group is discovered, the system proceeds to evaluate the subsequent rule levels, searching for a match group until either a match is found or the maximum allowed rule level for matching is reached.

• All attributes in a single matching rule level are connected using an AND relationship

  • When multiple attributes are present within a single rule level, they are interconnected by an AND relationship. During profile matching, all attribute values must align for profiles to be assigned to the same match group. For example, only when the values of all attributes are identical, profiles are considered as matches and grouped together for further processing.

• All attributes in an attribute type selector are connected with an OR relationship

  • When specifying attributes within the attribute type selector, attributes of the same type are linked through an OR relationship. For instance, consider the PhoneNumber Type where HomePhoneNumber and BusinessPhoneNumber are used. In this scenario, two profiles can be matched if either their HomePhoneNumber or BusinessPhoneNumber aligns. Consequently, the matching process allows for flexible matches based on either home or business phone numbers.

• The match result is eventually optimized.

  • Due to the near real-time nature of profile matching in the system, there is a possibility that a match group for your profile may be found at a lower (less optimized) rule level. Nevertheless, if a match is available at a higher (more optimized) rule level, the system will assign the profile to that particular group.

Note

When Identity Resolution performs rule-based matching, the order in which the rules that you have configured will be processed is dependant on how the data is ingested. For example, If you configure rules 1 and 2, rule 2 might be processed before rule 1. The processing order might change, but the end result will always be the same.

Automatic merging of similar profiles

After the profiles are matched, the Identity Resolution Job can optionally merge similar profiles based on the MaxAllowedRuleLevelForMerging you specify in the rule-based matching configuration.

Important

You cannot undo the consolidation process. It is recommended to turn on matching only first to evaluate the match result using the ListMatches and GetSimiliarProfiles APIs. You can turn on merging by setting the MaxAllowedRuleLevelForMerging using the UpdateDomain API.

Note

When merging two profiles, profile fields manually populated through an API call or the Agent Workspace will not be overwritten by profile fields automatically ingested from an integration or custom object type mapping.

For example, suppose a profile is created with FirstName “John” manually by an agent in the Agent Workspace. Another profile is created using an S3 integration with FirstName “Peter”. If these profiles are automatically merged, the FirstName “John” will be preserved.