Applying Center-out architecture to an agentic solution

Global Insurance Group (GIG) is a large healthcare insurance company that serves customers across multiple regions. GIG provides the following services:

• Policy enrollment
• Claims processing
• Provider network management
• Preauthorization requests
• Customer support

Over time, GIG has adopted several technology platforms, including cloud systems, a policy administration system, customer relationship management (CRM) applications, mainframe databases, and other legacy non-Pega systems.
GIG wants to introduce AI-powered customer interactions, improve the customer experience, increase operational efficiency, and modernize its digital architecture.
 

GIG serves customers through a customer portal, mobile app, call center, chatbot, and AI assistant. Each channel connects directly to systems of record, including policy administration, claims management, CRM, provider management, and member database systems.
The following diagram shows direct connections between each channel and multiple systems of record:

 Each channel often implements its own business rules, Validations, and integrations. This design duplicates logic and creates inconsistent customer experiences. As GIG adds digital and AI-powered channels, the number of channel-specific integrations increases.
A change to a business policy or system of record can require updates across several channels. These updates reduce agility and increase the risk of inconsistent outcomes for customers and employees.

Problem 1: Increasing complexity across channels
Each GIG channel contains separate business logic, Validations, and integrations with systems of record. This approach becomes difficult to maintain as GIG adds channels.
For example, if GIG introduces a voice assistant or a partner-facing AI agent, development teams must implement the same business functions again. This approach creates the following challenges:

• Duplicate development effort
• Inconsistent customer experiences
• Higher maintenance costs
• Slower delivery of new features

As the number of channels grows, GIG has more difficulty providing a consistent experience across customer interactions.

Problem 2: AI agents create new silos
GIG introduces an AI assistant that answers questions and helps customers complete common tasks. The AI assistant needs access to claims, policies, customer records, and other enterprise information.
If each AI agent implements separate integrations, business logic, and Decision Rules, the architecture creates the following challenges:

• Inconsistent responses across channels
• Duplicate business logic
• Increased governance and compliance risk
• Greater complexity when business policies or Rules change

Instead of simplifying the architecture, AI agents can become another set of siloed applications that are difficult to govern and maintain.

Problem 3: Tight coupling with systems of record
GIG channels connect directly to several systems of record. As a result, a change to one system can affect several customer-facing applications and AI channels.
For example, if GIG replaces its claims management system with a cloud-based solution, the change can require updates to the following channels:

• Customer portal
• Mobile app
• Call center tools
• Chatbot
• AI assistant

This design increases modernization costs and reduces the organization’s ability to respond to changing business needs. Additional point-to-point integrations also make the architecture less flexible and more difficult to maintain.

To address these challenges, GIG adopts a Center-out® Agent Architecture. Pega Platform™ becomes the central business orchestration layer between customer-facing channels and systems of record. The customer portal, mobile app, call center, chatbot, AI assistant, and future AI agents interact with Pega Platform instead of connecting directly to enterprise systems.

Pega Platform centralizes workflows, Decisions, business policies, and integrations. GIG can build a business service once and reuse it across channels. This approach supports consistent customer experiences and reduces duplicate logic and integrations.

Pega Platform manages the following components:

• Business workflows and Case Management
• Business Rules and policies
• Decision Management and AI governance
• Data access and integrations
• Agent APIs and Pega Digital Experience API (DX API)
• Security controls that support the organization’s compliance strategy

Pega Platform exposes governed enterprise services so that people and AI agents can use the same business logic without channel-specific implementations. 

For example, a customer can check a claim status by using the portal, mobile app, chatbot, call center, or AI assistant. Pega Platform processes each request through the same workflow and Decision logic, which supports consistent outcomes across channels.

The architecture also supports new channels. A voice assistant, partner portal, or autonomous AI agent can use business services that GIG already exposes through Pega Platform. When a system of record changes, GIG can update one integration in Pega Platform without modifying each customer-facing channel.

The following diagram shows Pega Platform as the orchestration layer between channels and systems of record:

The Center-out Agent Architecture helps GIG support consistent human and AI interactions, strengthen governance, accelerate delivery, and reduce application complexity.

Solution 1: Reusable business services across channels
Pega Platform centralizes workflows, Rules, and Decisions. GIG builds each business service once and reuses it across the customer portal, mobile app, call center, chatbot, and AI assistant.
For example, GIG creates the Submit Insurance Claim workflow as a reusable business service. Customers use the same workflow, Validations, and Business Rules regardless of the channel that they choose. This design supports a consistent experience and reduces development and maintenance effort.

Solution 2: Governed AI agent interactions
The Center-out Agent Architecture helps AI agents use governed business services instead of implementing separate logic and integrations. Pega Platform orchestrates workflows, policies, Decisions, and data access for each request.
For example, when a customer asks an AI assistant, “Why is my claim still pending?” the assistant invokes a business service that GIG exposes through Pega Platform. Pega Platform retrieves the required information, applies Business Rules and policies, and returns a governed response. The AI assistant follows the same governance, compliance strategy, and Decision standards as other channels. This approach removes duplicate logic and reduces implementation complexity.

Solution 3: Decoupled channels and systems of record
Pega Platform provides data virtualization and orchestration between channels and systems of record. Connectors and Data Pages access data from several systems while shielding channels from changes to the underlying technology.
For example, if GIG replaces its claims management system with a cloud-based solution, GIG updates the integration in Pega Platform. The portal, mobile app, call center, chatbot, and AI assistant continue to use the same business service and do not require separate integration changes.
This design reduces the effects of technology changes, supports modernization, and helps GIG add systems and channels with less disruption to the overall architecture.

Check your knowledge with the following interaction: