Pega Everywhere: The Center-Out Architecture Demo Playbook
18 Pega channel and integration patterns — from Web Portal and Salesforce to React, MCP, A2A, Copilot, and LangGraph — connected to one intelligent Pega core.
What Is Pega Center-Out Architecture in Practice?
The playbook demonstrates how one centralized Pega case, workflow, business rules, decisioning, data, AI, security, and governance layer can support multiple user experiences and external systems without duplicating enterprise logic.
The implementations include Pega Web Portal, Email Bot, Web Embed, Conversational Agent, Mobile Channel, Digital Messaging, Salesforce Extender, React applications using the Pega DX API, external AI agents using MCP, Pega-to-Pega Agent-to-Agent orchestration, Microsoft Copilot, LangGraph, Knowledge Buddy, Pega MCP Server, and Blueprint MCP Server.
Across every implementation, the architectural principle stays the same:
The same case. The same workflow. The same decisioning. The same governance. Only the channel changes.
This reference is built for Pega System Architects, Senior System Architects, Lead System Architects, enterprise architects, integration architects, AI architects, developers, and business technology leaders evaluating omnichannel and agentic AI architectures.
What This Playbook Demonstrates
- Centralized intelligence — Pega owns the case, workflow, rules, decisioning, AI, and governance
- Omnichannel delivery — the same business capability is available through out-of-the-box and custom channels
- Maximum reuse — business logic is built once instead of rebuilt for every interface
- Open integration — React, Salesforce, Copilot, and LangGraph connect through DX API, MCP, and A2A patterns
- Enterprise governance — external channels and AI agents delegate governed execution back to Pega
- Continuous expansion — new channels and demonstrations are added as Pega capabilities evolve
Read first: Pega Center-Out Architecture — Build Once, Deploy Anywhere The architectural case for why Pega centralizes enterprise intelligence at the core, and why every pattern below is just a different entry point into the same brain. Read the full article →
Pega Out-of-the-Box Center-Out Channels
| Status | Channel | Summary | Article / Demo | Source |
|---|---|---|---|---|
 |
Web Portal | Agentic AI embedded directly into the standard Pega web portal case lifecycle, guiding the case worker with grounded, governed recommendations at each step. | Smart Cases, Faster Decisions: Co-Pilot Your Workflow with Pega GenAI Agent Steps | Pega GenAI Cookbook |
|
Email Bot | An AI agent reads inbound emails, determines intent, and automatically creates Pega cases and triggers workflows. | Pega Agentic AI Email Bot: Intelligent Responses and Case Creation from Email | Pega GenAI Cookbook |
|
Web Embed | Embeds Pega Constellation case experiences directly into any external website using HTML5 Web Components, regardless of front-end framework. | Build Faster with Pega Web Embed | Constellation 101 |
|
Conversational Agent | A GenAI assistant embedded in the self-service portal guides applicants through a case step by step in natural language. | Agentic AI in Self-Service Portals: Guide Users to Complete Cases Seamlessly | Pega GenAI Cookbook |
 |
Mobile Channel | Native Pega mobile channel running the same case and workflow as every other channel. | Coming soon | — |
|
Digital Messaging | Case interaction via messaging platforms, driven by the same Pega workflow and decisioning. | Coming soon | — |
|
Salesforce Extender | Pega Process Extender embedded in Salesforce Lightning — Pega still owns the case lifecycle, business rules, and assignments behind the scenes. | Unlocking the Power of Pega Within Salesforce Lightning in Constellation | Constellation 101 |
|
Agent-to-Agent (Pega-to-Pega) | Pega Agentic Fabric orchestrates agents, workflows, data, and decisions across Pega applications, combining MCP connectivity with policy-based governance. | Pega Agentic Fabric: The Enterprise Brain for Agentic AI (MCP + A2A + Knowledge Buddy) | Pega GenAI Cookbook |
|
MCP Server | Pega Infinity '26 introduces native MCP Server support, exposing case types, workflows, and agents as discoverable tools for any MCP-compatible client. | Pega Infinity '26 MCP Server and Pega 25.1.3 MCP Connect | Pega Forums (standalone) |
|
Blueprint MCP Server | Positions Pega Blueprint as an MCP-callable design assistant, generating case types, data models, and UI specs from natural language. | Pega Blueprint MCP Server: Powering End-to-End Enterprise Orchestration | Pega GenAI Cookbook |
|
Knowledge Buddy | A governed semantic retrieval layer that supplies relevant enterprise knowledge to Pega and external AI agents without passing entire documents to the model. | Pega Knowledge Buddy: The Intelligence Layer Between Enterprise Knowledge and Agentic AI | Pega GenAI Cookbook |
Custom Applications and External AI Integrations with Pega
| Status | Pattern | Summary | Article / Demo | Source |
|---|---|---|---|---|
|
React JS Conversational Chat | A React front end delegates all case creation, knowledge retrieval, and reasoning to Pega Agentic AI via DX Agentic APIs. | Pega as the Agentic Brain: Powering React-Based Self-Service with Agentic AI | Pega GenAI Cookbook |
|
React JS with DX API | The REST API contract for managing AI agent conversations in Pega — the backbone that lets a React (or any external) app power a custom agentic UI. | Demo | — |
|
External AI Agent (via MCP) | Pega orchestrates specialized external agents — Gemini for analysis, Claude for risk and compliance — via MCP, while retaining identity, fraud, and governance control. | Agent Synthesis with Pega MCP: Orchestrating Enterprise AI Agents | Pega GenAI Cookbook |
|
Custom Android Applications | A native Android app calling into the same Pega case and workflow layer as every other channel. | Demo | — |
|
API-based Agent-to-Agent Integration | Pega’s A2A protocol orchestrates Google Gemini as the conversational front end and Anthropic Claude for deep analysis, each handing off execution to Pega. | Powering Enterprise Automation with Pega, Gemini, and Claude | Pega GenAI Cookbook |
|
Copilot to Pega (A2A) | Microsoft Copilot connects to Pega through A2A orchestration, turning natural-language questions into trusted, governed actions. | Copilot Working with Pega via A2A Orchestration | Pega GenAI Cookbook |
|
LangGraph to Pega | LangGraph’s Python-native stateful agent graphs connect to Pega case workflows via A2A, enabling self-healing multi-agent orchestration with Pega governing the audit trail. | Pega + LangGraph: Self-Healing Agentic AI using A2A | Pega GenAI Cookbook |
How Pega Center-Out Architecture Enables Reuse
Look across every row in both tables and one thing stays constant: Pega owns the case, the workflow, the decisioning, and the governance. The channel — whether it’s a browser, an inbox, a Salesforce page, or an external AI agent — is just where the interaction starts.
That’s the entire premise of Center-Out Architecture in practice: build the business capability once inside Pega, and let every current and future channel plug into it without rebuilding a thing.
Frequently Asked Questions
What is Pega Center-Out Architecture? Pega Center-Out Architecture places case management, workflows, business rules, decisioning, data access, AI, and governance at the center of the enterprise. Channels and external applications connect to this centralized intelligence instead of recreating business logic independently.
What is the Pega Everywhere Demo Playbook? Pega Everywhere is a continuously updated collection of Pega Center-Out Architecture demonstrations. It maps Pega out-of-the-box channels and custom integrations to practical articles, proof-of-concepts, and demonstrations.
Which Pega channels are included? The playbook includes Web Portal, Email Bot, Web Embed, Conversational Agent, Mobile Channel, Digital Messaging, Salesforce Extender, Knowledge Buddy, and other Pega capabilities.
Which custom applications can connect to Pega? Examples include React applications, native mobile applications, Microsoft Copilot, LangGraph, and external AI agents. These experiences connect to Pega through DX API, MCP, REST APIs, and Agent-to-Agent integration.
How does Pega integrate with external AI agents? Pega can expose or consume governed capabilities using the Model Context Protocol and Agent-to-Agent integration patterns. External agents provide specialized reasoning while Pega retains workflow orchestration, business rules, case execution, security, and auditability.
How does Center-Out Architecture prevent duplicated business logic? Business logic stays centralized in Pega. A web portal, mobile app, Salesforce interface, React application, or AI agent consumes the same workflows and decisions rather than implementing separate versions per channel.
Who should use this playbook? Pega architects, developers, enterprise architects, integration architects, AI architects, and business leaders evaluating omnichannel delivery, reusable workflows, and governed agentic AI.
Official Pega Resources
- Center-out Business Architecture — Pega
- Pega Center-out Architecture — Pega Academy
- Relevant official documentation for DX API, MCP, Agentic Fabric, and Constellation (see docs.pega.com)
About This Playbook
This playbook was created from a series of practical Pega proof-of-concepts covering out-of-the-box channels, custom React experiences, DX API, MCP, Agent-to-Agent orchestration, Salesforce, Microsoft Copilot, LangGraph, and external AI agents. Each linked entry points to its corresponding architecture, implementation guidance, or demonstration.
Share your recommendation — this playbook is updated as new POCs and Pega capabilities become available.
One Platform. Unlimited Possibilities.