Medium: a 10-prompt co-writer workflow instead of copy-pasting AI

Published on Medium in June 2026, this piece by Lawrence Clinton addresses a specific failure mode in AI-assisted writing: treating AI as a text dispenser rather than a collaborative tool. The article’s premise is that copying raw AI output and publishing it degrades search performance over time and produces content that reads like no one wrote it. The proposed alternative is a structured workflow in which AI participates at every stage, but humans shape the outcome.

Context: the problem being solved

Clinton’s starting point is a common one. Writers use AI to generate draft text, paste it into a document, edit lightly, and publish. The results look acceptable but perform poorly over time as search algorithms get better at identifying AI-generated patterns and readers get better at sensing when no one is really behind the words. The workflow described in the article is a structural response to this problem, not a set of tips.

The four-phase workflow

The ten prompts are organized into four phases. The first phase — research and strategy — covers competitor gap analysis, audience mapping, and headline generation. The second phase — writing and creation — covers first draft generation, producing analogies and concrete examples, and FAQ optimization. The third phase — editing and optimization — covers filler removal and internal link planning. The fourth phase — repurposing — covers social media adaptation and key takeaway extraction.

The critical distinction from copy-paste workflows is that each phase requires the writer to evaluate the output before moving to the next prompt. The AI is generating material to be worked with, not text to be used directly. Clinton describes this as using AI “as a highly efficient co-writer,” with fact-checking, expertise insertion, and voice refinement as non-negotiable before publication.

Key takeaway

The workflow is explicitly SEO-oriented and the audience is content creators and digital marketers rather than journalists or literary writers. Clinton claims the approach saves roughly half of total content creation time when the phases are followed consistently. The time savings come from having structured starting material at each stage rather than generating from scratch, not from reducing the number of decisions a writer makes.

The article is most useful for writers who are already using AI tools in some form and have noticed that raw AI output is not delivering results, and who want a more systematic approach rather than individual prompting tips.