Skip to main content
Version: V11

Agents

What is an Agent?​

In VIDIZMO, an Agent is a specialized AI assistant that extends the capabilities of VIDIZMO Agentic AI. Agents provide a customizable interface for users to interact with your organization’s knowledge, data, and automated workflows. While workflows handle the behind-the-scenes logic, Agents define how AI interacts with users, including behavior, branding, system prompts, and access to knowledge sources.

Agents are configured with system prompts, knowledge bases, branding, workflows, and user context, ensuring AI interactions are intelligent, consistent, and tailored to your organization.

How Agents work in VIDIZMO​

Agents in VIDIZMO transforms raw workflow outputs into consistent, branded, and context-aware interactions. The interaction flow can be broken down systematically:

  1. Receive User Input: The agent accepts natural-language requests from users via the AI chatbot interface.

  2. Apply Behavior: System prompts are applied to guide how the agent interprets and responds to user input. System prompts define the agent's tone, behavior, and response style.

  3. Manage Knowledge Access: The agent determines which knowledge sources to use based on the configured knowledge base criteria (content filters, categories, and search scope) and whether citations are included.

  4. Route to Workflow: Requests are forwarded to the connected workflow, where processing logic executes, including data retrieval, decision-making, and response generation.

  5. Format and Return Response: Once the workflow produces a result, the agent formats the response, ensuring it aligns with the configured tone, style, and interaction rules before delivering it to the user.

Key Points

  • The agent layer separates user experience from backend logic, making workflows reusable across different Agents.

  • Agents control behavior and personality, ensuring consistent interactions.

  • They mediate knowledge access, enabling precise, context-aware responses.

In summary, the agent governs the entire user-facing experience, transforming workflow intelligence into a guided, interactive, and personalized conversation.

How an Agent Relates to Workflows​

You can think of VIDIZMO's Agentic AI system as two tightly integrated parts:

  • Agent (the experience layer): Applies system prompts, knowledge-base settings, and routing configurations before sending requests to the workflow.

  • Workflow (the processing engine): Defines how data is processed, what actions are taken, and how responses are generated.

In this model, the agent defines the persona and presentation users interact with, while the workflow functions as the logic. Each agent connects to exactly one workflow. However, multiple agents can share the same workflow while maintaining different system prompts and knowledge base configurations.

Why Agents Matter​

Agents enable organizations to tailor AI interactions to specific business needs. By configuring system prompts, knowledge sources, suggested prompts, branding, and workflow routing, Agents provide:

  • Enhanced Productivity: Agents automate repetitive tasks such as responding to portal queries, guiding users through content, and routing requests to the correct workflows. This allows teams to focus on higher-value activities.

  • Controlled Knowledge Access: Agents manage which knowledge sources are used, including portal content or custom directories, ensuring accurate and context-aware responses.

  • Operational Efficiency: By routing requests intelligently and handling user interactions automatically, Agents reduce manual intervention and streamline processes across portals.

  • Adaptability and Learning: Agents can be reconfigured over time by updating workflows, system prompts, and knowledge base criteria, enabling continuous improvement in responses and task execution through manual adjustments.

Agents in the VIDIZMO Portal​

Agents are available to portal users and can be customized at both the portal and global levels:

  • Portal Agents: Configured within a specific portal for tailored user experiences.

  • Global Agent: A global agent configured at the portal level (tenant). Each portal can have one global agent that serves as the fallback when no specific agent is selected.

Extending Agents Beyond the Portal​

VIDIZMO also provides an Agent Widget Embed Code, allowing you to embed the agent as a chat widget in external websites or applications. This extends workflow powered AI interactions beyond the portal without additional development.

Summary​

In VIDIZMO, an Agent is a specialized AI assistant that extends Agentic AI and provides a customizable interface for interacting with your organization’s knowledge, data, and workflows. Agents automate tasks, offer insights, and connect to knowledge sources to support decision-making.tailored experiences.

Next steps​

Ready to build Agents?

Did you find this article helpful?