Agentic APIs

Profile Data and Enrichment

Follow the full profile pipeline from consent through enrichment and activation.

This page belongs to the older Agentic APIs lane: A2A, MuleSoft, browser-proxy, and profile-enrichment flows. The separate Kai lane covers PKM, consent, REST, and MCP.

This page explains how brand teams build a consent-first profile pipeline using MuleSoft agentic APIs and optional Plaid signals.

The data pipeline

1) Consent and identity

Collect the minimum identifiers required for enrichment and request user consent.

2) Profile creation or query

Create a new profile or retrieve an existing one from Supabase.

3) Public and Gemini enrichment

Generate a complete profile with public data and high-confidence gap filling.

4) Brand preference normalization

Translate signals into likes, dislikes, and brand affinities for CRM activation.

5) Activation

Store the merged profile in your CRM, CDP, or personalization system.

MuleSoft flow

MuleSoft orchestrates the agent chain, applies retry logic, and normalizes responses through DataWeave before data is stored or activated.

Plaid signals (optional)

Plaid is used to enrich merchant and category signals that strengthen brand affinity modeling.

warning

Plaid calls must be server to server. Do not run Plaid endpoints from the browser.

Common Plaid endpoints in this flow:

POST /link/token/create
POST /item/public_token/exchange
POST /accounts/get
POST /transactions/sync
POST /identity/get (if identity is consented)

Map Plaid outputs into brand affinity fields before calling the Brand agent.

Testable endpoints

Supabase Profile Creation Agent

Create a Supabase profile

Supabase Profile Update Agent

Update an existing Supabase profile

Supabase Profile Query Agent

Query a Supabase profile by phone or email

OpenAI Public Data Agent

Public data enrichment

Gemini Public Data Agent

Second-pass enrichment

Brand User Data Query Agent

Normalize likes, dislikes, and brand affinities

Quickstart prompts (paste into the Api Tester)

Use these prompts in the text field for the endpoints above.

Create a profile (Supabase Profile Creation Agent)

Create a profile for Sundar Pichai with email sundar.pichai@example.com, phone +1 6505559001, occupation CEO, city Mountain View, and country USA.

Update a profile (Supabase Profile Update Agent)

Update the city for Sundar Pichai to Mountain View and set the occupation to CEO.

Query an existing profile (Supabase Profile Query Agent)

Fetch all details for Sundar Pichai with phone +1 6505559001.

Enrich with OpenAI Public Data Agent

Provide a detailed JSON profile for Sundar Pichai, email sundar.pichai@example.com, phone +1 6505559001.

Enrich with Gemini Public Data Agent

Using the Gemini Public Data Agent, provide a detailed JSON profile for Sundar Pichai with email sundar.pichai@example.com and phone +1 6505559001, including demographics, lifestyle, and intent fields.

Normalize brand affinities (Brand User Data Query Agent)

Summarize likes, dislikes, and brand affinity for Sundar Pichai using the consented profile.

Example merged profile output

{
  "fullName": "Sundar Pichai",
  "email": "sundar.pichai@example.com",
  "phone": "+1 6505559001",
  "likes": ["fine leather goods", "limited drops", "Paris fashion week"],
  "dislikes": ["fast fashion", "discount-only messaging"],
  "brand_affinities": ["Dior", "Louis Vuitton", "Sephora"],
  "shopping_preferences": {
    "category_focus": ["handbags", "beauty", "fragrance"],
    "price_sensitivity": "low",
    "channel": "in-store and VIP events"
  },
  "confidence": 0.86
}

PreviousAgent Reference NextUse Cases