ISSUE NO. 06

Turn Repeated Conversations
Into Systems.

Stop solving the same problem again.

Look for repeated context, repeated decisions, and repeated output. Then turn the pattern into infrastructure.

ADAM JUKES FOUNDER, ONE BLADES

READ THIS IF YOU ONLY HAVE TWO MINUTES

06

If you keep having
the same conversation with AI,
stop starting again.

Turn the repeated work into a system.

ONE has done this repeatedly.

01

Instead of answering the same sizing question forever, build a sizing system.

02

Instead of repeatedly explaining setup, build setup guidance.

03

Instead of creating every publication from nothing, build a publication engine.

04

Instead of manually creating every assessment document, build an Assessment Builder.

Look for repetition in three places:

01

Repeated context

What do you keep explaining?

02

Repeated decisions

What do you keep evaluating using the same rules?

03

Repeated output

What do you keep creating from the same structure?

Then document the system.

Let AI help maintain it.

When the process changes, give AI the current Markdown file, talk through the change using voice, let it identify affected sections, then generate the new version.

Use AI to reduce
repeated thinking,
not just speed up
repeated typing.

THE WHOLE IDEA

When the same work keeps returning, look for the system hiding underneath it.

END OF QUICK VERSION CONTINUE FOR THE FULL PAPER ↓

Stop solving the task.
Start designing
the system.

The first stage of AI use feels magical because it saves time on individual tasks.

Write this.

Summarize that.

Give me ideas.

The next stage is more valuable.

You start noticing that many of those tasks are versions of the same underlying process.

That is when you should stop solving the task and start designing the system.

01

A real ONE example:
sizing.

If customers repeatedly ask which EDGE size they need, there are two ways to respond.

The first is to answer every person manually.

The second is to understand the recurring decision and build a sizing experience that helps the customer reach the answer themselves.

OPTION 01

Answer every person.

The first uses labor.

OPTION 02

Build the sizing experience.

The second creates infrastructure.

The same principle applies to AI work.

02

A real ONE example:
publication.

If every Skating CEO or ONE publication starts with a blank page and a completely new design process, the system is fragile.

A publication engine defines:

01 Canvas
02 Typography
03 Grid
04 Layout rules
05 Logo rules
06 Rhythm
07 Build order

Now the content changes while the frame remains stable.

AI can work inside the system rather than reinventing it.

The content changes.
The frame remains stable.

03

A real ONE example:
assessments.

The assessment idea follows the same logic.

Do not receive random emails with random video links and manually reconstruct what the skater is trying to submit.

Create a structured Assessment Builder.

01 Requirements

The skater sees what is required.

02 Information

They fill in the relevant information.

03 Structured output

The output arrives in a consistent format.

04 Human assessment

The coach spends more time assessing and less time organizing.

That is what good systems do.

They protect the human work
that actually matters.

04

Find the
repetition.

Ask yourself three questions.

01

What context do I keep repeating?

Move it into a project source file.

02

What decisions do I keep making?

Turn the criteria into a framework.

03

What output do I keep producing?

Create a reusable structure or builder.

Repeated context becomes a source.
Repeated decisions become a framework.
Repeated output becomes a system.

05

Let AI maintain
the system.

This is important enough to repeat.

Do not build a doctrine and then turn yourself into the person responsible for manually editing it forever.

Suppose a creator policy changes.

Open the relevant AI project.

Provide the current Markdown file.

Then speak:

This is what changed and why.

Read the current document and tell me every place this affects.

Ask me anything you need before updating it.

The AI inspects the implications.

You answer.

Then:

Create the complete new version.

Do not give me editing notes.

Give me the replacement .md file.

That is system maintenance through conversation.

06

Voice makes maintenance
easier.

This is another reason voice-to-text matters.

When a system has changed in several subtle ways, describing those changes aloud is often much faster than manually editing multiple sections.

You can say:

We stopped doing this because of X.

This new rule replaces it.

The terminology has changed.

This section still applies.

This other part is now wrong.

AI can then map that natural explanation back onto the structured document.

07

How this translates
outside ONE.

The same principle works almost anywhere repeated knowledge work appears.

01

A teacher can systemize lesson planning.

02

A freelancer can systemize proposals.

03

A manager can systemize weekly reviews.

04

A family can systemize travel planning.

05

A coach can systemize athlete assessments.

06

A creator can systemize content production.

The point is not automation for its own sake.

The point is to remove repeated low-value work so human attention can go toward the work that still needs judgment.

When the same problem
keeps coming back,
stop solving it.

Build the thing that solves it.

THINK WITH AI · ISSUE NO. 06

Use AI to reduce
repeated thinking,
not just speed up
repeated typing.

When the same problem keeps coming back, build the thing that solves it.

NEXT ISSUE 07 / 07

ISSUE NO. 07

Keep the Human
In Charge.