TThe Diary of a CEO
← All frameworks
Innovation

The AI Augment-or-Automate Decision

Use AI to cut cost or multiply a person's competitive output

Difficulty
Moderate
Time to result
~weeks to results
Steps
5
Confidence
94%

Hanauer frames AI adoption as a competitive choice rather than an automatic headcount exercise. A company can automate a task to lower cost, or it can keep the person, equip them with AI, and try to outperform competitors through greater output or quality. Priestley gives examples from his companies: AI-supported marketing created more appointments, which led to more sales hiring, while an AI layer supplied context, reports, and conversation prompts. The framework compares both routes against the same customer outcome. It also preserves uncertainty: the speakers acknowledge that AI may still reduce hiring through attrition and that its rapid distribution differs from earlier technologies. The decision therefore requires measured results rather than a blanket claim that AI creates or destroys jobs.

Origin

Extracted from The Diary of a CEO

Core principles

  • 01A business can compete by being cheaper or by being better
  • 02Automation removes tasks while augmentation expands useful output
  • 03The decision should be judged against competitive outcomes, not tool novelty

How to run it

  1. 1

    Define the competitive outcome

    Specify whether the business needs lower cost, higher quality, faster delivery, or more demand.

    Watch out Do not begin with a tool and search for a reason to use it.

  2. 2

    Model automation

    Estimate what happens if AI performs the task with less human involvement, including quality and oversight costs.

    Watch out A task saving is not a customer outcome.

  3. 3

    Model augmentation

    Estimate what the same person could deliver with appropriate AI tools, context, and training.

    Pro tip Look for downstream demand that greater output may create.

  4. 4

    Run a bounded comparison

    Pilot both plausible routes and compare cost, quality, throughput, demand, and errors.

    Watch out Near-term results may not predict labour effects as models improve.

  5. 5

    Choose and revisit

    Adopt the stronger route for the current workflow and reassess as capabilities and market conditions change.

In the wild

More appointments, then more salespeople

Priestley says his companies used AI in marketing and appointment setting. Increased appointment volume then created demand for people to conduct the final sales conversations.

AI augmented the funnel and was associated with additional hiring in the described companies.

Common mistakes

Assuming automation is the only gain

Removing labour may sacrifice a larger opportunity to improve quality or output with the same people.

Claiming the labour effect in advance

The episode presents competing possibilities and does not establish a universal net employment result.

Is it for you?

Best for

It is best for teams deciding how to deploy AI in an existing workflow.

Not ideal for

It is not ideal where output quality, safety, or accountability cannot be measured reliably.

From the transcript

There are two ways to compete. One is to be cheaper. The other is to be better.

Nick Hanauer · 1:03:30

You could keep that person, give them the right tools, and outcompete your competitors.

Nick Hanauer · 1:04:30

From the episode

Death of the Middle Class: Billionaire vs Entrepreneur DEBATE - Daniel Priestley v Nick Hanauer