Generative AI Leader · Free Practice Question Medium
Question 27
A company is designing a customer service agent. For some common, straightforward queries like "What are your store hours?", they want the agent to follow a predefined conversational flow with fixed responses to ensure absolute accuracy and consistency. However, for more complex or novel inquiries, they want the agent to leverage a large language model to understand intent and generate more flexible, natural-sounding responses.
Which type of agent architecture best describes this approach?
-
A
Purely Generative Agent
-
B
Hybrid Agent
-
C
Unsupervised Agent
-
D
Purely Deterministic Agent
Reveal correct answer
Correct answer: B
Explanation
A hybrid agent combines the predictability of deterministic flows with the flexibility of generative models.
INCORRECT: Purely Deterministic Agent
A purely deterministic agent would only follow predefined paths and would not be able to handle novel inquiries with flexible, generated responses.INCORRECT: Purely Generative Agent
A purely generative agent would use an LLM for all responses, which might not guarantee the absolute accuracy and consistency desired for simple, factual queries like store hours.CORRECT: Hybrid Agent
A hybrid agent architecture combines the strengths of both deterministic (rule-based or state machine-based) and generative (LLM-based) approaches. It uses deterministic flows for predictable, straightforward interactions where accuracy is paramount, and leverages generative capabilities for more complex, nuanced, or open-ended conversations. This matches the company's requirements.INCORRECT: Unsupervised Agent
"Unsupervised Agent" is not a standard classification for conversational agent architectures in this context. Agent types usually refer to how they process information and generate responses (deterministic, generative, hybrid).
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
