---
id: "framework-difference-analysis"
type: "framework"
source_timestamps: ["§ Step 3. Analyze the differences between your initial view (from Step 1) and AI's output."]
tags: ["evaluation", "diagnostics"]
related: ["framework-four-step-ai-development", "concept-looks-right-but-isnt"]
steps: ["\\\"Identify what AI added that you missed (e.g.", "new angles", "broader scope", "overlooked options).\\\"", "\\\"Identify what AI got wrong or missed entirely (e.g.", "stale data", "wrong assumptions", "missing context).\\\"", "\\\"Identify what looks right but isn't (e.g.", "plausible", "well-structured outputs that fail due to subtle domain-specific nuances).\\\""]
sources: ["reskilling"]
sourceVaultSlug: "hbr-seg-reskilling"
originDay: 10
articleStem: "hbr-edu-32-help-employees-get-better-with-ai"
sourceUrl: "https://hbr.org/2026/06/help-employees-get-better-not-just-faster-with-ai"
sourceTitle: "Help Employees Get Better—Not Just Faster—with AI"
---
# AI Output Difference Analysis

A diagnostic tool used in [[framework-four-step-ai-development|Step 3]] to categorize the *delta* between a human's initial hypothesis (Step 1) and the AI's generated output, so the professional can synthesize the best of both — learning from the machine's broader scope and from their own contextual superiority. Three buckets:

1. **What AI added that you missed** — new angles, broader scope, overlooked options.
2. **What AI got wrong or missed entirely** — stale data, wrong assumptions, missing context.
3. **What [[concept-looks-right-but-isnt|looks right but isn't]]** — plausible, well-structured outputs that fail on subtle, domain-specific nuances.

Executing bucket 3 depends on [[prereq-domain-knowledge|underlying domain knowledge]] and non-public context. The output of this analysis feeds directly into the [[concept-reasoning-trail|reasoning trail]] produced in Step 4.