Clapper: A content creator's five-step approach to AI-assisted writing in 2026
Harri Drake, writing on the Clapper blog in June 2026, argues that the most reliable way to use AI in a writing workflow is to identify the parts that are logistical and separate them from the parts that are irreducibly personal. The article is structured around five specific practices and stays grounded in how a working content creator actually moves through a piece rather than describing AI tools in the abstract.
The five practices
Draft generation. Drake uses AI to produce an initial version of a piece, then edits extensively until the voice feels authentic. The emphasis is on editing rather than accepting output: the AI draft is a starting point that removes blank-page friction, not a finished product that gets minor corrections.
Voice documentation. Before prompting for any piece, Drake maintains a reference document containing characteristic phrases, typical angles, and personal opinions on recurring topics. This gets fed into prompts to keep AI-generated drafts closer to the writer’s actual register rather than the model’s default tone.
Filtered ideation. For any piece, Drake asks the AI to generate multiple angles on the topic, then selects personally which ones are worth developing. The generation step is faster with AI; the selection step requires human judgment that can’t be outsourced without losing the editorial character of the content.
Content repurposing. Existing published pieces get converted into different formats — a long article becomes a series of shorter social posts, a video script becomes a newsletter section — using AI to handle the structural transformation while the writer edits for each new context.
Fact-checking. Every AI-assisted piece goes through manual verification of statistics and claims before publication. Drake treats this as non-negotiable regardless of how confident the AI output sounds.
Tools and context
Drake uses ChatGPT for caption drafting and brainstorming, Claude for longer-form writing where voice consistency across paragraphs matters more, and Gemini for tasks that benefit from Google Workspace integration. The three-tool split reflects different strengths rather than a preference for any single model.
The piece was written for content creators producing regular written and video output — people who need AI to accelerate production without producing content that reads or sounds unlike them.
The central argument
Drake’s framing is that authentic human-centered content has become a competitive advantage as AI-generated material becomes easier to recognize and receives lower engagement. The practical implication is that using AI well in 2026 means protecting the parts of the work that create that advantage while letting AI carry the parts that don’t require it.
Who this is useful for
Writers and content creators who have experimented with AI tools and found the output generic, and who want a framework for getting more use out of AI without giving up what makes their work distinct. The five-step structure is practical enough to adapt to different content types — newsletters, articles, scripts — and the voice documentation practice in particular applies to anyone who has a recognizable personal style they want to maintain at scale.