Skip to main content
Version: Current

IB-X Agent Model

Overview​

The IB-X Agent Model defines how automation is expressed, structured, and executed within the IB-X platform.

Rather than exposing multiple automation technologies or runtimes, IB-X uses Agents as the primary abstraction for all automation scenarios. This ensures consistency across design, execution, governance, and scaling.

What is an Agent?​

An Agent in IB-X is a self-contained unit of automation that encapsulates:

  • Execution logic and behavior
  • Integration and service access
  • Execution context and state
  • Governance and lifecycle metadata

All agents in IB-X follow a common lifecycle and governance model and are executed using the same agent framework.   When automation requires UI-level access to local systems, execution is delegated to Local Agent components running within the customer network, while design, control, and monitoring remain centralized.

Logical Agent Classification​

IB-X does not impose hard or physical distinctions between different types of agents.

All agents in IB-X share:

  • A common execution runtime
  • A unified lifecycle
  • Centralized governance through the AI Command Center

Agents are logically classified based on the activities and capabilities they use during execution.

For example:

  • An agent that uses deterministic automation activities is commonly referred to as an Automation Agent
  • An agent that uses conversational or AI-enabled activities is commonly referred to as a Conversational Agent

This classification is conceptual and descriptive, not structural, allowing agents to evolve organically over time.

Common Agent Patterns​

The following patterns represent commonly observed ways agents are built and used within IB-X.

Automation Agent Pattern​

Agents following this pattern typically:

  • Execute deterministic workflows
  • Perform business process automation
  • Integrate systems and applications
  • Process and validate data
  • Manage long-running or stateful processes

Automation Agents may execute:

  • Centrally within the IB-X platform, or
  • As Local Agents within the customer network when UI-level access is required

Local Agents​

Local Agents execute within customer environments using:

  • Smart Station
  • Smart Buddy
  • Assistant Buddy (optional / evolving)

Execution remains centrally governed by the AI Command Center.

Conversational Agent Pattern​

Agents following this pattern exhibit interaction-driven behavior, in which natural language input guides automation execution.

In this pattern, an Agent can:

  • Engage users through text- or voice-based interactions
  • Interpret user intent and maintain conversational context
  • Determine the next action to perform at runtime
  • Invoke tools dynamically based on intent, context, and business rules

Tools invoked by a conversational agent may include:

  • Individual activities
  • Sub-workflows or reusable automation flows
  • External system integrations

Conversational behavior does not represent a separate agent type.   An Agent follows this pattern when it leverages conversational or AI-enabled activities as part of its execution, while continuing to operate within the same runtime, lifecycle, and governance model as all other agents.

Hybrid Automation​

IB-X enables hybrid automation scenarios, where:

  • Conversational capabilities determine what actions are required
  • Deterministic automation determines how those actions are executed

Because agent classification is logical rather than physical, a single agent may evolve to support both patterns over time.

Agent Lifecycle & Governance​

All agents in IB-X follow a consistent lifecycle:

  1. Design
  2. Versioning
  3. Deployment
  4. Execution
  5. Monitoring
  6. Governance and policy enforcement

The AI Command Center centrally manages this lifecycle for all agents, including Local Agents executing within customer environments.

Summary​

The IB-X Agent Model:

  • Establishes Agents as the foundational abstraction for automation
  • Uses logical classification rather than physical agent types
  • Supports deterministic, conversational, and hybrid automation
  • Ensures consistency across execution, governance, and scaling

This model provides a flexible and future-proof foundation for all automation built on IB-X.