An editorial stack is the set of tools and steps that take a piece of content from idea to published. Adding AI to an editorial stack can cut production time significantly — but only if you add it to the right stages. Add it in the wrong place and you get generic content faster, which isn't an improvement.
The Editorial Stack Concept
Every piece of content goes through stages, whether or not they're formalised:
- Ideation: What topics are worth covering and why
- Briefing: What the piece needs to achieve and cover
- Drafting: Getting words on the page
- Editing: Making those words work
- Fact-checking: Verifying claims
- Publishing: Preparing and distributing
AI is useful at some of these stages and actively harmful at others. The skill is knowing which.
Where AI Adds Value
Ideation — useful, with caveats
AI is good at generating topic ideas from a seed: a keyword, a question your customers ask, a trend you've noticed. What it can't do is tell you which topics will resonate with your specific audience — that requires the market knowledge you have and it doesn't.
Use AI to generate twenty options. Use your judgment to select two. This is faster than generating from scratch and better than using AI's judgment about what's interesting.
Briefing — very useful
Once you know what you're writing, use AI to generate a draft brief: key points to cover, potential angles, what the piece should accomplish. A good brief from AI takes three minutes to generate and three minutes to edit. That's faster than writing a brief from scratch and better than no brief at all.
Drafting — useful for the first 70%
AI is excellent at getting you from blank page to rough draft. The draft will be structurally competent and tonally generic. Your job is to edit it into something specific, opinionated, and distinctly yours. If you publish AI drafts without editing, readers notice — not because they know AI wrote it, but because nothing distinctive is said.
Editing — limited usefulness
AI can check grammar, suggest clearer phrasing, and flag passive voice. It can't tell you whether your argument is compelling, whether your examples are the right examples, or whether your conclusion follows from your evidence. Those are editorial judgments that require a human reader with domain knowledge.
Fact-checking — do not delegate to AI
AI is the source of facts that need checking, not the checker. Asking an AI to verify its own claims is circular. Use primary sources for fact-checking.
Building Your AI-Assisted Editorial Process
A practical implementation:
- Monday (15 min): Use AI to generate ten topic ideas based on this week's themes. Select two. Write a brief for each using the AI, edited by you.
- Tuesday–Wednesday: Draft with AI. Edit aggressively. Add examples, specific details, and a clear point of view that wasn't in the AI draft.
- Thursday: Fact-check specific claims manually. Run a final read for tone consistency.
- Friday: Prepare for publishing. Use AI for SEO meta descriptions and social copy (both are short, low-stakes tasks where AI drafts work well).
The Quality Trap
The risk of AI-assisted content is producing more of the same: competent, pleasant, and indistinguishable from everything else in the genre. Avoid this by ensuring every piece has at least one element that AI couldn't have provided: a specific experience, an original observation, a proprietary data point, or a genuinely held opinion. That element is what readers remember. AI can scaffold everything else.