E01¶


Research	R0021 — Prompt engineering definitions
Run	2026-03-25
Query	Q003
Source	SRC04
Evidence	SRC04-E01
Type	Factual

Microsoft's prompt engineering guidance and explicit acknowledgment of its non-engineering nature

URL: https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/prompt-engineering

Extract¶

Microsoft states: "In practice, the prompt acts to help the model complete the desired task, but it's more of an art than a science, often requiring experience and intuition to craft a successful prompt."

Key recommendations: 1. State the role and expected outcome — Subjective. 2. List topics to avoid — Structural. 3. Specify format plainly — Subjective ("plainly"). 4. Use grounding data for factual scenarios — Structural/architectural. 5. Validate responses — Process recommendation. Microsoft warns: "even when using prompt engineering effectively you still need to validate the responses."

Quantifiable recommendations found: 0. Microsoft explicitly frames prompting as art, not engineering.

Relevance to Hypotheses¶

Hypothesis	Relationship	Strength
H1	Contradicts	Zero quantifiable recommendations; vendor explicitly calls it "art"
H2	Supports	"More art than science" is a vendor's own characterization
H3	N/A	No quantifiable elements to create a mixed picture

Context¶

Microsoft's explicit statement that prompting is "more of an art than a science" is notable because it comes from a vendor selling prompt engineering services through Azure. This is a vendor acknowledging the non-engineering nature of its own recommended practice.