Skip to main content

​
Custom Properties

​
Table of Contents

  1. Understanding Properties
  2. Accessing Properties Settings
  3. Property Types
  4. Creating a New Property
  5. Property Configuration Options
  6. Scope: Contacts vs Conversations
  7. AI Extraction with Prompts
  8. Managing Existing Properties
  9. Using Properties in the Platform
  10. Best Practices for Properties

​
Understanding Properties

Properties are custom data fields that extend the information you can track on contacts and conversations.

​
What Properties Enable

Business-Specific Data: Track information unique to your operations that is not covered by standard fields. Segmentation: Categorize contacts and conversations for filtering and analysis. AI Extraction: Automatically capture information from conversations. Workflow Support: Enable processes based on property values.

​
Property Examples

For Contacts:
  • Customer status (Lead, Prospect, Customer)
  • Source (Website, Referral, Campaign)
  • Interest level (Hot, Warm, Cold)
  • Assigned account manager
  • Product interests
For Conversations:
  • Call purpose (Sales, Support, Inquiry)
  • Sentiment (Positive, Neutral, Negative)
  • Issue type (Billing, Technical, General)
  • Follow-up required (Yes/No)
  • Sale amount

​
Accessing Properties Settings

Navigate to the properties configuration page.
  1. Click Settings in the main navigation
  2. Select Properties from the settings menu
  3. The Properties settings page displays

​
Page Layout

The properties page shows:
  • Header with title and create button
  • Table listing all defined properties
  • Information about each property
  • Action menus for management

​
Property Types

Properties support different value types to match your data needs.

​
String (Text)

Free-form text values:
  • Names, descriptions, notes
  • Open-ended information
  • No predefined options
Example: Custom notes field

​
Number

Numeric values:
  • Scores, amounts, quantities
  • Mathematical comparisons possible
  • Decimal or integer
Example: Lead score, order amount

​
Boolean

True/false values:
  • Yes/No questions
  • Toggle states
  • Binary choices
Example: VIP customer, requires callback

​
Date

Calendar date values:
  • Important dates
  • Deadlines
  • Scheduled events
Example: Next follow-up date, contract expiration

​
Enum (Options)

Selection from predefined choices:
  • Consistent categorization
  • Controlled vocabulary
  • Dropdown selection
Example: Status (Open, Pending, Closed)

​
Creating a New Property

Define a new custom property for your organization.

​
Accessing Creation

  1. Navigate to Settings > Properties
  2. Click New Property button
  3. Property creation dialog opens

​
Required Fields

Identifier: A unique name for the property:
  • Used as the internal reference
  • Appears in column headers
  • Should be concise but descriptive
  • No spaces or special characters recommended
Example identifiers: customer_status, lead_source, interest_level Scope: Where this property applies:
  • Contact: Attached to contact records
  • Conversation: Attached to conversation records
Value Type: Select the type of data:
  • String, Number, Boolean, Date, or Enum

​
Optional Configuration

Default Value: A preset value for new records:
  • Applied when property is not explicitly set
  • Provides consistent baseline
Options (for Enum type): Define allowed values:
  • Add each option
  • Order as they should appear
  • Edit or remove as needed
Prompt (for AI extraction): Instructions for extracting values from conversations.

​
Property Configuration Options

Fine-tune property behavior with additional settings.

​
Default Value

Set what value appears when not specified:
  • For Status fields, might be “New”
  • For Boolean, might be “false”
  • Provides consistent starting point

​
Allowed Values (Enum)

For dropdown/enum properties, define the choices: Adding options:
  1. Enter the option text
  2. Add to the list
  3. Repeat for all choices
  4. Order can be adjusted
Example options for Status: “Lead”, “Qualified”, “Opportunity”, “Customer”, “Churned”

​
Fallback Value

If AI extraction fails or value is unclear:
  • Fallback provides a safe default
  • Prevents empty values
  • Maintains data consistency

​
Scope: Contacts vs Conversations

Properties can apply to either contacts or conversations.

​
Contact Properties

Apply to individual contact records:
  • Persist across multiple conversations
  • Describe the contact themselves
  • Updated through conversations or manually
When to use: Information about who the person is. Examples:
  • Customer type
  • Preferred language
  • Account tier
  • Industry

​
Conversation Properties

Apply to individual conversation records:
  • Specific to that interaction
  • May vary between calls with same contact
  • Often extracted during the call
When to use: Information about what happened in the call. Examples:
  • Call outcome
  • Topics discussed
  • Sentiment detected
  • Next actions

​
Choosing the Right Scope

Ask yourself:
  • Does this describe the person or the interaction?
  • Should it persist across conversations?
  • Is it likely to change each call?

​
AI Extraction with Prompts

Configure properties to be automatically extracted from conversations.

​
How AI Extraction Works

  1. Conversation occurs and is transcribed
  2. AI analyzes the transcript
  3. Based on your prompt, AI identifies the property value
  4. Value is automatically set on the record

​
Writing Extraction Prompts

The prompt tells the AI what to look for: For enum properties: “Determine the caller’s primary interest. Options are: ‘Product A’, ‘Product B’, ‘General Information’, ‘Support’. Select the best match based on what they asked about.” For boolean properties: “Did the caller express interest in scheduling a demo? Answer true if they explicitly requested a demo or meeting, false otherwise.” For string properties: “Extract the specific product or service the caller mentioned they currently use.”

​
Prompt Best Practices

Be specific: Clearly describe what to look for. Include context: Explain what the values mean. Reference options: For enum, list all valid choices. Handle uncertainty: Specify what to do if unclear.

​
Extraction Limitations

AI extraction works best when:
  • Transcript is clear and accurate
  • Information was explicitly discussed
  • Prompt provides clear guidance
May be less reliable when:
  • Audio quality was poor
  • Information was only implied
  • Conversation was very short

​
Managing Existing Properties

Maintain and modify properties over time.

​
Viewing Property Details

The properties table shows:
  • Identifier (name)
  • Scope (Contact or Conversation)
  • Value type
  • Allowed values (if enum)
  • Default value
  • Creation date

​
Editing Properties

Modify property configuration:
  1. Click the actions menu for the property
  2. Select Edit
  3. Modify desired settings
  4. Save changes
Editable fields:
  • Display settings
  • Default values
  • Allowed options (for enum)
  • Extraction prompt
Note: Changing type may affect existing data.

​
Deleting Properties

Remove properties no longer needed:
  1. Click the actions menu
  2. Select Delete
  3. Confirm the deletion
Considerations:
  • Existing values may be lost
  • Filter options using this property will break
  • Reports may be affected

​
Using Properties in the Platform

Properties integrate throughout the platform.

​
In Tables and Lists

Column display: Add property columns to contact and conversation tables. Filtering: Filter records by property values. Sorting: Some properties can be used for sorting.

​
In Detail Pages

Viewing: Property values display in detail sections. Editing: Modify values directly on records. History: Changes may be tracked in activity timelines.

​
In Analytics

Segmentation: Group data by property values. Comparison: Compare metrics across property segments. Trends: Track property value distributions over time.

​
With AI Employees

Context: Properties can inform AI employee responses. Extraction: AI employees populate property values during calls. Personalization: Future conversations can use property data.

​
Best Practices for Properties

​
Planning Your Property Schema

Start with use cases: What questions do you need to answer? Avoid redundancy: Do not create overlapping properties. Think about reporting: How will you use this data? Consider growth: Leave room for future needs.

​
Naming Conventions

Be consistent: Use the same style across properties. Be descriptive: Names should be self-explanatory. Keep it concise: Long names are unwieldy. Avoid abbreviations: Unless universally understood.

​
Value Standardization

Use enum where possible: Consistent values are more useful than free text. Define clear options: Each choice should be distinct. Include “Other” carefully: Can become a catch-all. Update as needed: Add new options when patterns emerge.

​
Regular Maintenance

Review usage: Are all properties being used? Clean up unused: Remove properties that add no value. Update options: Add new enum values as business evolves. Check extraction accuracy: Verify AI is extracting correctly.

​
Extraction Prompt Quality

Test your prompts: Review extracted values for accuracy. Iterate: Refine prompts based on results. Be explicit: Vague prompts produce inconsistent results. Update for changes: Modify prompts when business context changes.

​
Summary

Custom properties transform your platform from a generic tool into one tailored to your specific business needs. By thoughtfully defining properties that capture meaningful information, you enable better segmentation, more insightful reporting, and automated data capture through AI extraction. Invest time in planning your property schema, maintain it actively, and leverage properties throughout the platform for maximum value.