AI Copilot Value-Curve Model
Delegate lower-value tasks to AI so people can concentrate on judgment and creativity.
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 94%
The AI Copilot Value-Curve Model treats a job as an array of tasks rather than a single indivisible role. Teams identify repetitive, lower-judgment activities that a copilot can perform alongside a person, such as research, rewriting, or preparing drafts. The collaborative framing aligns incentives: the copilot helps the worker achieve the organization's goal without presenting itself as an immediate replacement. The human reviews outputs, teaches the system through corrections, and redirects saved time toward strategy, creativity, relationships, and final decisions. As reliability improves, more tasks can be delegated, allowing the person to move progressively higher up the value curve while retaining accountability for the result.
Origin
Extracted from Marketing Against the Grain during Nicholas Holland and Kieran Flanagan's discussion of copilots, task decomposition, and job redesign.
Core principles
- 01Treat every role as a collection of tasks.
- 02Use collaboration to make adoption less threatening.
- 03Delegate repetitive work before judgment-intensive work.
- 04Measure success by increased human value, not removed headcount.
- 05Reinvest saved time in creativity, strategy, and decision-making.
How to run it
- 1
Decompose the Role
List the recurring tasks that make up the person's job. Separate production activities from decisions, relationships, and creative judgment.
Pro tip Observe a real workweek instead of relying only on a formal job description.
Watch out Do not assume every visible activity contributes equal value.
- 2
Rank Tasks for Delegation
Score tasks by repetition, risk, required context, and need for human judgment. Begin with repetitive, reversible, and easily reviewed work.
Pro tip Rewriting, summarization, initial research, and draft preparation are useful starting points.
Watch out Do not delegate consequential decisions merely because the model can generate an answer.
- 3
Give the Copilot a Shared Goal
Explain the outcome, audience, constraints, and quality standard. Position the copilot as a collaborator helping the person succeed.
Pro tip Include examples of acceptable and unacceptable work.
Watch out A bare command without context encourages generic output.
- 4
Review and Train
Have the person inspect the copilot's work, correct errors, and capture recurring feedback in instructions or reference material. Continue until performance is predictable.
Pro tip Batch similar outputs so reviewers can detect patterns.
Watch out Silent correction prevents the workflow from improving.
- 5
Reinvest the Capacity
Use the time saved for tasks requiring stronger ideas, editing, strategy, or customer understanding. Evaluate whether the person's total contribution has increased.
Pro tip Track both hours saved and improvements in quality or business outcomes.
Watch out Filling recovered time with more low-value activity defeats the purpose.
In the wild
A social media specialist gives a copilot routine research, rewriting, monitoring, and draft preparation. The specialist reviews the material and spends the recovered time developing stronger points of view, improving platform fit, and interacting meaningfully with the audience.
→ The marketer produces better work at greater scale while retaining creative and strategic control.
A creator uses AI to produce rough drafts, transformations, and supporting research. Instead of increasing the number of generic posts, the creator spends more time on ideation and editing, which remain the hardest and most valuable parts of the process.
→ Automation increases the quality of the final content rather than merely its volume.
Common mistakes
Automating the Whole Job at Once
Roles contain tasks with very different risk and judgment requirements. Delegate selectively rather than treating the role as one automatable unit.
Measuring Only Labor Reduction
A copilot can create more value through better research, ideas, and decisions even when no role disappears. Measure increased capability as well as efficiency.
Keeping the Repetitive Work
If workers spend saved time producing more low-quality output, they never move up the value curve. Protect time for judgment-intensive work.
Is it for you?
Best for
It is best for knowledge workers who spend substantial time on repeatable research, drafting, formatting, or administrative tasks.
Not ideal for
It is not ideal for roles whose tasks are all highly sensitive, novel, or dependent on unrecorded human context.
From the transcript
“all jobs are just the array of tasks and you'll have to like figure out the task that you can give to the co-pilot”
“the co-pilot's job is not to replace the human it's to basically make the human more valuable”
“that will allow us now to move higher up the value curve”
From the episode
Using GPT Agents For Content Creation (Tools & Predictions) ft Nicholas Holland
Nicholas Holland