The AI Content Stack
Match each writing task to AI based on how much differentiation it needs.
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 5
- Confidence
- 94%
The AI Content Stack routes writing work according to how much originality and differentiation the output requires. At the lower end, internal communications such as memos and proposals are consensus-oriented, so AI can generate useful drafts or synthesize information. Functional content such as basic landing pages and product explanations also benefits, especially when clarity matters more than distinction. At the top of the stack, audience-building content depends on point of view, humor, voice, and memorable packaging. AI can support research, brainstorming, and outlining there, but a human should create or substantially rewrite the final version. The mechanism preserves human effort for the work where distinctiveness produces the greatest return.
Origin
Extracted from Marketing Against The Grain during Kieran Flanagan and Kipp Bodnar's mid-year assessment of AI writing tools.
Core principles
- 01AI performs best on consensus writing tasks.
- 02Differentiated content still requires human judgment and voice.
- 03Research, outlining, and first drafts offer immediate efficiency gains.
- 04The more audience impact a piece needs, the less suitable raw AI output becomes.
How to run it
- 1
Classify the writing task
Determine whether the task is internal communication, functional content, or differentiated audience-building content.
Pro tip Judge the task by the level of originality its audience expects, not merely by its format.
Watch out Do not assume every public-facing asset belongs in the same risk category.
- 2
Identify the consensus requirement
Ask whether the output mainly needs to summarize facts, organize information, or communicate a standard answer clearly.
Pro tip AI is particularly useful when several reasonable writers would produce similar answers.
Watch out A technically correct answer can still damage positioning if it sounds interchangeable with competitors.
- 3
Assign AI an appropriate role
Let AI draft routine communication, assist with functional content, and provide research or outlines for differentiated work.
Pro tip Give the model source material and explicit constraints rather than relying on a generic prompt.
Watch out Do not treat the first generated draft as publication-ready.
- 4
Add human differentiation
For high-impact content, introduce original argument, experience, humor, voice, and message discipline.
Pro tip Use AI output as raw material while preserving the thinking involved in writing.
Watch out Skipping this step produces average, consensus-driven content.
- 5
Review against the intended outcome
Check whether the final piece merely communicates information or actually earns attention, trust, and recall.
Pro tip Compare it with competitor content and remove phrases that anyone in the category could say.
Watch out Efficiency is not a win if the result weakens the brand.
In the wild
A marketer uses AI to synthesize weekly notes into action items, drafts a basic feature landing page with human editing, and uses AI only for research and outlining on a provocative LinkedIn post. The marketer writes the post's final argument and voice personally.
→ Routine work ships faster while the public thought-leadership piece remains distinctive.
A manager collects feedback from several people, asks ChatGPT to categorize it, synthesize recurring themes, and produce key takeaways, then compares the result with a manual analysis.
→ The manager reduces review time and identifies where human interpretation may have introduced bias.
Common mistakes
Publishing the average
Generic model output reflects consensus language and can make a brand sound indistinguishable from everyone else.
Rejecting useful assistance
Avoiding AI entirely sacrifices valuable support for research, synthesis, outlining, and routine drafts.
Confusing correctness with quality
Accurate functional copy may still lack the voice and insight required to attract an audience.
Is it for you?
Best for
Marketing teams deciding where AI writing can save time without weakening their brand.
Not ideal for
Teams seeking a fully autonomous system for producing original, high-end creative work.
From the transcript
“Again, when we think about writing, there's internal comms, external, functional writing, which is to me is like how do I just get landing pages…”
“AI is great at consensus, right? It's trying to get to a consensus answer.”
“But I don't think it's like the top echelon of what writing is about.”
From the episode
The State Of A.I. In 2023: Useful Tools Vs Hype (#120)