# Gap-Analysis Prompting Is a Distinct Skill From Generation Prompting ## Claim Asking AI to find what's missing or wrong in existing work is a different, underexplored skill from asking AI to generate new work outright. <!-- atomicity: 1 idea ✓ --> ## Scope & conditions - Holds when the creator already has a body of prior work and a defined target audience to gap-check against. - Breaks down for someone starting from a blank page with no existing asset or persona research to interrogate — gap-analysis prompting needs something to find the gaps *in*. - Does not claim gap-analysis prompting is inherently better than generation prompting in general; it claims the two require different skills and produce different failure modes when confused. ## Evidence Christopher Penn's workflow for repurposing a practitioner-facing newsletter into an executive-facing one deliberately routes AI into interrogation rather than drafting: personas built from ensembled model research, a scraped Reddit corpus loaded into NotebookLM, and Claude Code querying that corpus roughly 60 times to generate hard questions Penn then answers himself, by voice, while driving. He states the underlying principle directly: "If you use AI to automate your outputs, you're likely to get weaker... if you use AI to automate your inputs, you're likely to get stronger." ## Links - [[AI & the Future of Knowledge Work MOC]] — an underexplored use of models that fits knowledge work better than generation does - [[3. Resources/Permanent Notes/Delegating a Decision to AI Does Not Transfer Accountability for It]] — because extends; both notes locate the durable human role (accountability, in one case; the actual drafting and thinking, in the other) as the part AI shouldn't be handed. - [[3. Resources/Permanent Notes/Scaled Content Is Defined by Its Production Method Not Its Quality]] — because contrasts; that note is about how content gets classified by production method regardless of output quality, while this one argues a specific production method (AI-for-interrogation, human-for-drafting) produces a qualitatively different creator outcome, not just a different content classification. - [[3. Resources/MOCs/DevRel MOC]] — applies-in; mining real support-forum or community threads for blind spots before rewriting documentation is the direct developer-content analog to Penn's Reddit-mining step. ## Tests / falsifiers - Compare self-assessed problem-solving capability over a multi-month period between creators who primarily use AI to interrogate/gap-find versus creators who primarily use AI to draft finished output. - Run a blind reader study comparing perceived authenticity/voice-fidelity of content produced via gap-analysis prompting against content fully AI-drafted from the same source material. ## Maintenance - Last tended: 2026-08-16 - Next review: 2026-11-16 ## Source - [[3. Resources/Literature Notes/Almost Timely News How To Expand and Improve Content with AI, Part 1]] — reports the personas-Reddit-NotebookLM-Claude Code-voice-dictation pipeline and the inputs-vs-outputs thesis.