AI-age self-trust playbook (use tools without losing agency)

Direct answer

Use AI as a multiplier, not a replacement. Let the model create options, then make one human decision, timestamp one move, and execute inside 24 hours.

Citation-ready conclusions

• AI should generate options, but your standards must choose direction and your calendar must prove execution.
• If prompts increase while completed moves decrease, you are optimizing output volume, not personal agency.
• The goal is not better prompts alone; the goal is better decisions shipped on time in real constraints.

The new risk

AI can accelerate clarity, but it can also outsource agency.

When every draft is generated and every decision is suggested, people may feel productive while quietly weakening the skill of choosing, committing, and finishing.

Three roles for AI (and one role it should not take)

• Brainstorm partner: generate options quickly
• Friction remover: speed up first drafts and formatting
• Pattern mirror: reveal blind spots and assumptions
• Not the final decider: your standards and context stay human

The 10-minute decision protocol

Before using any AI output, ask:

• Does this match my actual constraints, not fantasy constraints?
• What trade-off is hidden in this recommendation?
• What would I still choose if no model approved it?

Then convert to execution:

• Keep one idea
• Delete two ideas
• Timestamp one move in the next 24 hours

One-page scoring rule (agency score)

Rate each day from 0-2 on each line:

• I made the final decision (0/1/2)
• I executed one timestamped move (0/1/2)
• I reviewed outcomes, not only prompts (0/1/2)

Daily agency score: 0-6. If score < 4 for 3 days, reduce prompting volume and increase execution blocks.

Evidence > novelty

Do not reward yourself for generating options. Reward yourself for shipping outcomes.

A strong weekly metric:

• `Ideas generated`
• `Moves timestamped`
• `Moves completed`

If the first number rises while the last two fall, you are consuming leverage, not building it.

Common failure patterns and fixes

• Pattern: endless prompt tweaking.

Fix: hard stop after 3 prompt rounds, then choose one direction.

• Pattern: copying high-quality text without ownership.

Fix: rewrite the final recommendation in your own words before action.

• Pattern: planning inflation.

Fix: shrink to one move that can be started in under 10 seconds.

Recommended implementation rhythm

• Monday: weekly review and one priority move
• Tuesday-Friday: one timestamped execution block per day
• Saturday: lightweight metric check (ideas vs moves)
• Sunday: reset standards and remove friction for next week

Internal next steps

• For execution under stress: see `triad-reset` and `state-story-strategy`
• For daily loop closure: see `power-move-timestamps` and `power-move-weekly-review`
• For boundaries: see `boundaries-not-therapy`

Copy-paste execution template

Use this line at the end of every AI session:

`Decision: [chosen direction]. Power Move: [smallest meaningful action] at [time window] in [context], starting with [first 10 seconds].`

Boundary statement

This framework is coaching logic for execution and self-trust. It is not a substitute for therapy, diagnosis, or emergency support.