Entity detection

The Entity Detection feature provides named-entity recognition support for your conversations. The Entity Detection feature identifies and extracts critical data from your conversations. Use the feature to:

  • Simplify the way your organization detects entities across numerous conversations.
  • Locate PII, PCI and PHI data that complies with PCI SSC, HIPAA, GDPR and CCPA.
  • Create custom entities to support product features and workflows in your organization.

Entities can be detected in recorded conversations processed using the Async API, or during the course of a real-time conversation using the Streaming API. Entities are detected by default when you use the Async API.

When you detect entities in a recorded or real-time conversation, the response includes managed entities provided by Symbl.ai and custom entities that you create with the Management API.


Entities use cases

Entities can be used in a variety of ways:

  • Automate workflows: Route a conversation or parts of a conversation to different individuals based on the entities detected. Automatically populate forms, such as a customer name, address, phone number, and appointment date and time from a sales call.
  • Recommendations: Identify entities such as events in user conversations and then make contextually-appropriate recommendations to that user. For example, if the mention of a holiday is detected, recommend relevant holiday events or products to the user.
  • Organize conversations: Extract entities and apply the detected entities to conversations as metadata, which can be used to create relevant groups of conversations.
  • Monitoring trends: Monitor social media trends, and extract entities from social media forums like Twitter and Reddit. Extract companies, people, and places mentioned on social media to regulate content.
  • Gather customer feedback: Analyze feedback for products, services, and people. Identify competitors mentioned in customer conversations, and questions related to particular products and services.

Managed entities

The following list includes some examples of the managed entities provided by Symbl.ai:

  • Personal Identifiable Information (PII) entities: Names, ages, birthdays, social security numbers, and drivers license numbers.
  • Payment Card Industry (PCI) entities: Bank accounts, routing numbers, credit card numbers, expiration dates, and credit card verification values (CVVs).
  • Protected Health Information (PHI) entities: Health conditions, blood groups, injuries, and medical statistics.
  • General entities: Events, file names, times, and URLs.

This page includes a complete table of supported entities.

Supported entities

The Entity Detection feature supports entity formats for the following countries:

  • United States of America
  • Canada
  • India
  • United Kingdom
  • Australia

The following table is a complete list of the supported entities for each class of data.

PIIPCIPHIGeneral
NameBank_AccountBlood_TypeEvent
Name_GivenMoneyName_Medical_ProfessionalFile_Name
Name_FamilyCredit_CardOrganization_Medical_FacilityTime
AgeCredit_Card_ExpirationConditionURL
Phone_NumberCVVDrugNumerical_PII
Email_AddressRouting_NumberInjuryDate
Passport_NumberMedical_MiscDate_Interval
SSNMedical_ProcessDuration
Drivers_LicenseStatistics
DOBHealthcare_Number
LocationDose
Location_Address
Location_Zip
Location_City
Location_State
Location_Country
Location_Coordinates
Marital_Status
IP_Address
Language
Occupation
Origin
Username
Password
Gender_Sexuality
Physical_Attribute
Religion
Political_Affiliation
VIN
Organization
Zodiac_Sign
Account_Number
Vehicle_ID

Each entity has a type and subType. In the preceding table, the column headings identify the type of each managed entity. The rows identify the subType of each entity.


Custom entities

Use the Management API to create and manage custom entities. Custom entities identify and extract data that exactly matches vocabulary that you provide. Custom entities can be used to build implementations that aren't covered by Symbl.ai's managed entities, such as detecting organization-specific vocabulary like brand names or internal departments.


What’s next?

The following pages describe the operations you can perform with bookmarks: