Which design choice MOST reduces the risk of hallucinations?

The correct answer is C. Strong grounding in Salesforce data. Hallucinations occur when an LLM generates plausible-sounding but factually incorrect responses, typically because it has no reliable source of truth. Grounding-connecting the agent's responses to verified, real-time Salesforce data through Data Cloud, object queries, or retrieva

AI Features for Service (e.g., Einstein Bots, Next Best Action)

Question

Options

ALarger prompts
BHigher temperature settings
CStrong grounding in Salesforce data
DOpen-ended instructions

How the community answered

(31 responses)

A
3% (1)
C
94% (29)
D
3% (1)

Explanation

Hallucinations occur when an LLM generates plausible-sounding but factually incorrect responses, typically because it has no reliable source of truth. Grounding-connecting the agent's responses to verified, real-time Salesforce data through Data Cloud, object queries, or retrieval-augmented generation-anchors the LLM's output to actual facts. This is the single most effective hallucination mitigation. Larger prompts (A) add context but can introduce noise and don't substitute for real data. Higher temperature settings (B) increase randomness and creativity, making hallucinations more likely, not less. Open-ended instructions (D) give the model too much freedom to fabricate, the opposite of what's needed.

Topics

#AI Hallucinations#Data Grounding#AI Design Principles#Reliable AI

Community Discussion

No community discussion yet for this question.

Full AI-201 Practice