Skip to main content

Data Types

Replace rigid forms with conversational data collection. Data Types define what information to capture, and AI assistants collect it naturally through dialogue—structured data without the friction.

What Data Types Do​

Data Types turn conversations into structured records:

  • Define schemas with typed fields and validation
  • Collect via conversation instead of form-filling
  • Search records with text matching or semantic similarity
  • Sync with external systems via API
  • Import/export CSV for bulk operations

Key Features​

Conversational Collection​

AI weaves data collection into natural dialogue—asking follow-up questions, handling corrections, and validating input conversationally.

Enable AI-powered search on text fields. Search "frustrated customer" finds records mentioning "angry" or "disappointed" even without exact matches.

External ID Sync​

Sync records with external systems (CRM, EHR) using external IDs. Upsert operations create or update records automatically.

Audit Trail​

Every record change is tracked—who changed it, when, and what values changed. Essential for compliance.

How It Works​

Schema Definition    →    AI Collection    →    Structured Records
(fields & types)         (conversational)      (queryable, exportable)

You define the data structure. When a member engages with a workflow, the AI collects the information naturally and saves it as structured records linked to that member.

Field Types​

TypeDescriptionExample
TextFree-form textName, notes
NumberNumeric valuesAge, score
BooleanYes/NoConsent, subscription
DateDate valuesBirth date
EmailValidated emailContact email
PhonePhone numberMobile
SelectSingle choiceStatus, category

Dynamic Field Options​

Make your data types smarter with conditional logic and custom validation.

Conditional Select Options​

Show or hide dropdown options based on other field values using visibleWhen. Useful when options depend on a previous selection.

Example: State-specific insurance plans

A "State" field with options like CA, TX, NY. An "Insurance Plan" select field where:

  • "Medi-Cal" only appears when State = CA
  • "CHIP Texas" only appears when State = TX
  • "Medicaid" appears for all states

Configure each option's visibleWhen condition:

state == "CA"

The AI will only offer relevant options based on what the member has already provided.

Custom Validation​

Add validation rules beyond basic field types using validateCel. The validation runs when a value is collected, and the AI will ask the member to correct invalid input.

Example: End date must be after start date

end_date > start_date

Example: Age must be 18 or older

age >= 18

Validation messages are shown conversationally—the AI explains what's wrong and asks for a corrected value.

For full syntax details, see the CEL Expression Guide.

Common Use Cases​

Use CaseData Captured
Patient IntakeDemographics, medical history, insurance
Lead QualificationCompany info, budget, timeline, pain points
Incident ReportsDate, location, severity, actions taken
Support TicketsIssue details, priority, resolution

Example: Patient Intake​

Schema:

  • Full Name (text, required)
  • Date of Birth (date, required)
  • Insurance ID (text)
  • Allergies (text)
  • Reason for Visit (text, required)

Conversation:

AI: "To get you scheduled, I'll need a few details. What's your full name?"

Member: "Sarah Johnson"

AI: "Thanks Sarah! And your date of birth?"

The AI collects each field naturally, validates as it goes, and saves a structured record—no form required.

Pricing & Limits​

Feature availability and limits are determined by your subscription plan. See the pricing page for details.