6.2 Explain AI Agent Design Principles, Autonomous Capabilities, and Orchestration AI-TECHNICAL-PRACTITIONER Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A financial firm deploys an AI agent to monitor market data and execute trades. The agent must react to price changes within milliseconds, but the firm's compliance policy requires that all trades be logged with a timestamp and rationale. Which autonomous capability is most critical for the agent to balance these requirements?

   • AA feedback loop that reviews past trade performance and adjusts future strategies
   • BDecision-making that selects the optimal trade strategy based on historical data analysis
   • CAction execution that is fast enough to capture market opportunities while recording the decision context for each trade
   • DPerception that continuously scans market data feeds and identifies price changes as they occur
   Reveal Solution Discussion Select an answer first
2. 7expert · hard

   A company is deploying an AI agent to manage its social media accounts. The agent must post content, respond to comments, and report on engagement metrics. The company wants the agent to be autonomous but also wants to ensure that the agent does not post content that violates brand guidelines. Which design approach best ensures brand compliance while maintaining autonomy?

   • AA separate compliance agent that reviews all posts after they are published and deletes any that violate guidelines
   • BA human approval step for every post, eliminating the agent's autonomy
   • CA feedback loop where the agent's proposed posts are checked against a brand-guideline rule set before posting, with violations blocked and flagged for human review
   • DA fully autonomous agent with no checks, relying on the agent's training to avoid brand violations
   Reveal Solution Discussion Select an answer first
4. 8expert · hard

   A large enterprise is deploying AI agents to handle IT support tickets. One agent triages tickets, another handles password resets, a third handles software installation requests, and a fourth escalates unresolved issues to human staff. The triage agent must route tickets to the correct specialist, but the specialists are sometimes overloaded. The enterprise wants to minimize human involvement while ensuring no ticket is dropped. Which orchestration design best balances these constraints?

   • AA static routing table where the triage agent always sends password resets to the password agent, software requests to the software agent, and so on, with no load awareness
   • BA workflow where the triage agent routes based on ticket type, and each specialist agent has a feedback loop to signal when it is overloaded, triggering the triage agent to hold or re-route the ticket
   • CA delegation model where the triage agent sends all tickets to a single generalist agent that handles everything
   • DA central orchestrator that assigns every ticket to a human first, and humans decide which agent should handle it
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   Which capability enables an AI agent to gather information from its environment to inform its decisions?

   • APerception
   • BAction execution
   • COrchestration
   • DDecision-making
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A software development team wants to use multiple AI agents to build a feature. One agent writes code, another reviews it for security vulnerabilities, and a third integrates the code into the main branch. The team wants the process to run without human intervention unless the security reviewer finds a critical issue. What orchestration pattern should they implement?

   • AA parallel execution model where all three agents work on the same code simultaneously and the team merges the results manually
   • BA sequential pipeline where the coding agent's output feeds the security reviewer, whose approval triggers the integration agent, with a conditional stop on critical findings
   • CA delegation model where the coding agent decides whether to invoke the security reviewer or the integration agent based on its own judgment
   • DA single autonomous agent that performs all three roles sequentially without any external coordination
   Reveal Solution Discussion Select an answer first