Tutorial: Adopt your first agent
This tutorial provides a step-by-step walkthrough for locating, adopting, and executing an AI agent within Rival Enterprise. Designed for team members and operational staff, this process takes approximately five minutes to complete.
You will select an enterprise-approved agent, clone it into your workspace, supply task context, manage human-in-the-loop checkpoints, and evaluate execution outputs.
Prerequisites
Section titled “Prerequisites”Before beginning this tutorial, verify that the following setup steps are complete:
- Workspace access — you are authenticated into your organization’s workspace via Single Sign-On (SSO).
- Approved assets — your enterprise administrator has approved at least one agent in the enterprise Marketplace.
- System connections — connected enterprise applications are linked if required by the target agent.
Walkthrough
Section titled “Walkthrough”-
Understand the role of an enterprise agent.
Before deploying an agent, it is essential to understand how agents differ from standard AI chatbots:
- Goal-directed execution — instead of providing static text answers to prompt queries, an agent is assigned an operational objective and coordinates tool interactions to achieve it.
- Integrated capabilities — agents use authorized enterprise connectors, specific functions, and contextual instructions to execute tasks directly across your corporate systems.
- Governed actions — agents operate within strict administrative permission boundaries and pause for human sign-off whenever gated actions are triggered.
-
Locate an approved agent in the Marketplace.
- Navigate to the Marketplace from the primary left navigation menu.
- Filter by category — select Agents to view the repository curated by your IT and governance administrators.
- Evaluate available solutions — select an agent tailored to a routine, high-frequency task within your team, such as lead enrichment, inventory monitoring, document extraction, or competitive research.
- Verify system requirements — review the agent details page to verify which connectors, functions, or authorization scopes are required for execution.
-
Adopt the agent into your workspace.
Click Adopt on the target agent detail card.
- Instance creation — adopting clones an independent instance of the agent configuration directly into your personal Workspace inventory.
- System authentication — if the agent requires access to connected systems that you have not yet authenticated, you will be prompted to authorize your credentials.
- Configuration integrity — base prompt instructions, tool links, and security guardrails established by your administration team are preserved during adoption.
-
Supply operational context and execute.
Open the newly adopted agent from your Workspace tab.
- Provide task input — initiate the agent by supplying clear operational parameters, background context, target documents, or explicit business objectives.
- Execute the task sequence — the agent analyzes your request, formulates an action sequence, calls authorized system tools, and processes data.
- Manage approval gates — if the agent attempts a gated system action, such as sending an external email, updating a CRM record, or modifying a financial ledger, execution pauses. Review the proposed action in your Requests queue and select Approve or Reject.
-
Review deliverables and refine outputs.
Once the agent completes its run:
- Inspect deliverables — review the generated work, report, or system updates directly within the execution viewer or in your connected target application.
- Refine iteratively — if the output requires adjustment, provide additional context or refined instructions and run the agent again.
- Audit telemetry — access the Activity log to view full execution details, including API calls made by the agent, time elapsed, and approval timestamps.