Fundamental Concepts

To effectively utilize the HCL UnO Agentic AI Builder, it's essential to learn the foundational elements that define its functionality. This section introduces the key terms and fundamental ideas that are crucial for designing, building, and deploying intelligent Agentic AI agent and their associated workflows. By familiarizing yourself with these core concepts, you will gain a deeper understanding of how the Agentic AI Builder operates and how to leverage its full potential for your automation needs.

Agents

  • Agents are intelligent systems capable of autonomously executing tasks - ranging from simple, rule-based workflows to complex, adaptive operations powered by AI.
  • In the UnO Agentic AI BuilderH, an agent is an autonomous, modular system that leverages a Large Language Model (LLM) to make decisions and perform tasks based on defined inputs, tools, logic, and memory.
  • Agents are constructed using configurable building blocks that define how they process information, interact with tools, and respond to dynamic conditions.
  • They can be designed to handle a wide range of tasks—from basic actions to complex, multi-step operations -depending on how their components are assembled.
  • Regardless of their complexity, all are treated uniformly as Agentic AI Agents within the system.
  • Using the visual builder, users can orchestrate agent behavior, define control logic, and integrate tools and memory - entirely without writing code.

Large Language Model (LLM)

  • A foundational AI model that interprets natural language input and generates intelligent output. Agents use LLMs to understand and respond to user queries.

Small Language Model (SLM)

  • A Small Language Model is a lightweight language model designed for efficiency and faster inference with fewer parameters compared to LLMs. SLMs are ideal for resource-constrained environments, edge devices, and domain-specific tasks, offering lower latency and reduced computational cost while maintaining reasonable accuracy.

Subagents

  • In the UnO Agentic AI Builder, subagents are modular, task-specific AI agents that function as part of a larger agent composition. They operate under the control of a parent or orchestrating agent to execute specific subtasks within a broader workflow.
  • Subagents are designed to enhance modularity, scalability, and debuggability in complex automation scenarios by isolating discrete responsibilities into reusable components.

    Key Characteristics:

    • Purpose-Built: Sub-agents are intended to perform a single task or a closely related set of tasks.

    • Scoped Visibility: By default, sub-agents are embedded within an agent definition and are not visible in the agent library (AI Agents catalogue).

    • Orchestrated Execution: They are invoked by a parent agent that controls the overall logic and flow.

Package

  • A Package is a bundled collection of agents grouped together for deployment or reuse. It allows teams to manage related agents as a single unit.

Templates

  • Templates provide predefined structures for building agents or workflows. They serve as reusable blueprints that streamline setup, enforce standards, and reduce manual configuration effort.
  • The UnO Agentic AI BuilderTemplates are preconfigured workflows that solve common business problems - like summarizing documents, forecasting demand, or triaging support tickets. These templates include structured prompts, tool access, and logic paths to help you deploy agents faster. You can select a template, modify input parameters, and deploy it to suit specific needs.

Tools and integrations

  • Tools are external functionalities or systems that an AI agent can invoke to perform specific actions or access data. The HCL UnO Agentic Builder allows you to define and integrate these tools, enabling agents to interact with:

    • Enterprise Applications: ERPs (for example, SAP), CRMs, HR systems.

    • Databases: SQL, NoSQL databases.

    • APIs: Custom or third-party APIs.

    • RPA Bots: Triggering robotic process automation tasks.

    • Cloud Services: Interacting with cloud-based platforms and services.

Workflow

  • The defined execution path of an agent or group of agents. It includes tasks, conditions, decision points, and the overall sequence of operations.