MMarketing Against The Grain
← All frameworks
Productivity

AI 80/20 Editor Model

Let AI create the first 80%, then spend human effort on the decisive 20%.

Difficulty
Easy
Time to result
~days to results
Steps
5
Confidence
95%

The AI 80/20 Editor Model divides work according to comparative advantage. AI rapidly produces the broad structure, alternatives, and initial language that would otherwise consume most of the operator's time. The human then concentrates on the smaller but more consequential portion: checking assumptions, simplifying language, injecting beliefs, applying domain expertise, choosing between alternatives, and ensuring the result fits the intended audience. The time saved on drafting should not simply shorten the task; it should improve the quality of the final editorial pass. This model applies to strategy, positioning, content, plans, and product briefs, but it requires an accountable reviewer. The 80/20 split is directional rather than a fixed measurement and should change when the task carries greater risk.

Origin

Extracted from Marketing Against The Grain during the hosts' review of an AI-generated go-to-market strategy.

Core principles

  • 01Use AI to compress the first-draft workload.
  • 02Reserve human attention for judgment, relevance, and differentiation.
  • 03Treat generated work as editable material, not final truth.
  • 04Reinvest saved time in the parts that affect outcomes most.
  • 05Ship only what a responsible human has reviewed.

How to run it

  1. 1

    Define the Quality Bar

    State the decision, audience, evidence requirements, and outcome the final work must support.

    Pro tip Write evaluation criteria before generating the draft.

    Watch out Without a quality bar, fluency can be mistaken for usefulness.

  2. 2

    Generate the Broad Draft

    Use AI to create the initial structure, analysis, options, and language.

    Pro tip Supply sufficient context so the first 80% is relevant rather than generic.

    Watch out Do not treat the percentage as proof that the draft is mostly correct.

  3. 3

    Locate the Critical 20%

    Identify claims, strategic choices, audience language, and creative decisions that will determine the result.

    Pro tip Spend most review time where an error would change the outcome.

    Watch out Cosmetic editing can distract from faulty assumptions.

  4. 4

    Apply Human Judgment

    Correct errors, add beliefs and experience, remove jargon, and choose recommendations you genuinely support.

    Pro tip Ask whether you would personally defend every major decision.

    Watch out Human review must be substantive, not ceremonial.

  5. 5

    Verify and Finalize

    Check the edited result against source evidence and the quality bar before using or publishing it.

    Pro tip Use external validation for consequential factual claims.

    Watch out AI plus an unverified human edit can still produce confident misinformation.

In the wild

Refining Cemetery Positioning

AI produces a broad market analysis and go-to-market plan. The marketer then simplifies the language for nontechnical cemetery managers, removes clever but distracting phrasing, strengthens the validation phase, and makes the value proposition directly about converting paper plot records into manageable digital records.

AI saves drafting time while human editing improves relevance and credibility.

Common mistakes

Rubber-Stamping the Draft

A quick read-through does not supply the judgment that the final 20% requires.

Spending the Savings Elsewhere

The model's value comes partly from reinvesting saved time into deeper refinement and verification.

Applying 80/20 Literally

High-risk work may require far more human review than the shorthand suggests.

Is it for you?

Best for

Knowledge and creative work where AI can draft quickly but context, taste, or strategic judgment determines quality.

Not ideal for

Tasks requiring independently verified precision where an approximate AI-generated first pass creates unacceptable risk.

From the transcript

And then you have to be like the editor, just because you got the idea browser prompt doesn't mean like don't accept it as like…

Greg Eisenberg · 10:00

I always think with AI AI is going to do 80% and that you do the last 20% but you have so much time to…

Host · 12:00

make sure that you believe in what you're building.

Greg Eisenberg · 10:00

From the episode

How to Start a $1M Business Using Only AI