Data Sources → Customer Touchpoints

Web / Mobile Apps

Customer-facing web and native mobile applications are typically the highest-volume source of behavioral events — every page view, product view, cart action, search, and click.

High-Level Design

Angular / native app → .NET Core microservices on Azure → downstream activation.

Touchpoint

Web / Mobile App

Angular SPA + native iOS/Android

→

Ingestion

.NET Core Ingestion API

Cxos.Ingestion.Client NuGet SDK → Azure API Management → AKS

→

Processing

.NET Core Stream Worker

Azure Event Hubs → Azure Container Apps

→

Foundation

Data Lakehouse

Azure Data Lake Storage Gen2 + Azure Purview

→

Intelligence

.NET Core Identity / Analytics API

Azure DB for PostgreSQL + Azure OpenAI

→

Activation

.NET Core Activation API

Azure Service Bus → personalization / CRM

💼 Business Context

Highest-volume, lowest-cost-to-instrument channel — usually the first source wired into a CXOS rollout
Directly drives conversion-rate optimization, personalization, and cart-abandonment recovery
Owned by Digital / Product / Growth; often the P&L justification for the platform investment

🔌 Technical Overview

The Web SDK (TypeScript) and Mobile SDK (Swift / Kotlin) are thin wrappers around the same contract published in the Cxos.Ingestion.Client NuGet package, so the browser, the mobile app, and the .NET Core Ingestion API all agree on one event schema. Client apps call Azure API Management, which authenticates the request (Azure AD B2C for end-user tokens), applies rate limiting, and routes to the Ingestion API — an ASP.NET Core Web API packaged as a Docker container and running on AKS. The API validates the payload and publishes it onto Azure Event Hubs for downstream processing.

Typical Events

page_view product_viewed add_to_cart search_performed app_opened checkout_started

💾 Sample Event Payload

{
  "event": "product_viewed",
  "event_id": "9f2c1e6a-2b41-4e9d-8f3a-1c7d9a0b6e2f",
  "timestamp": "2026-08-01T10:22:14Z",
  "anonymous_id": "a1b2c3d4",
  "user_id": "cust_004821",
  "context": {
    "channel": "web",
    "session_id": "sess_77af",
    "page_url": "/products/sku-1029",
    "device": { "type": "desktop", "os": "Windows", "browser": "Chrome" }
  },
  "properties": {
    "product_id": "sku-1029",
    "product_name": "Wireless Headphones",
    "price": 2499,
    "currency": "INR"
  }
}

🔗 Integration Points

Cxos.Ingestion.Client NuGet package (Azure Artifacts feed) — typed contracts + Polly retry policy
Azure API Management — gateway, auth, throttling
Ingestion API — ASP.NET Core microservice, packaged as a Docker container, running on AKS
Azure Event Hubs — Kafka-compatible ingestion backbone
Azure AD B2C — end-user authentication / consent context
Application Insights — end-to-end request tracing from the SDK through every downstream microservice

⚠️ Non-Functional Considerations

Scale: AKS autoscaling + Event Hubs throughput units absorb flash-sale traffic spikes without event loss
Latency: sub-200ms edge-to-ingest typical, needed for real-time personalization
Reliability: Mobile SDK queues events locally and retries via Polly when connectivity returns
Security/Privacy: Azure AD B2C consent claims gate SDK initialization; PII fields tagged in the shared contract

🎯 Enterprise Example

A retail enterprise runs its Ingestion API as an ASP.NET Core microservice on AKS, fronted by Azure API Management. During a flash sale, Azure Event Hubs absorbs a 12x traffic spike while the Stream Worker autoscales on Azure Container Apps. The Analytics API — sharing the same Cxos.Ingestion.Client contract — flags users who viewed a promoted SKU but didn't purchase within 10 minutes, and the Activation API pushes a personalization event through Azure Service Bus, lifting flash-sale conversion by 6.2%.

← Back to Customer Touchpoints