WORKFLOWS / PLACES TO BEGIN

Bring a task.
Build a method.

Clean up a dataset, compare sources, prepare a file for another app, or set up a routine. Copy a prompt, try it on sample files, then build on the result with a follow-up.

01 / CODING & DATA

From messy inputs
to a checked result.

Find the odd records, build a summary, and keep the cleanup steps ready for the next file.

EXAMPLE WORKFLOW
01 / ORDERED MATTER

01 / ORDERED MATTER

INSPECTTRANSFORMCHECK
INPUT

A few source files.

Some CSV or JSON samples and a question you want answered. Work on copies.

METHOD

Inspect. Transform. Check.

Work out the structure, say what is being assumed, and verify the totals and odd cases.

OUTPUT

A summary and its working.

A new result file, a short explanation, and the code that produced it.

EXAMPLE PROMPT
Inspect the sample files in this folder and describe their schemas. Identify inconsistent records before changing anything. Then create a separate summary file and explain how you checked it. Keep the original files unchanged.
Why the follow-up is faster

02 / RESEARCH & SYNTHESIS

Follow the evidence.
Keep the materials.

Build a comparison with references, flag claims that disagree, and update it when a new source turns up.

EXAMPLE WORKFLOW
02 / SOURCE FIELD

02 / SOURCE FIELD

SOURCESCOMPARISONREFERENCES
INPUT

A focused question.

A few source documents, what to compare them on, and how far any browsing should go.

METHOD

Organise the evidence.

Pull out comparable facts, note where each came from, and spot what conflicts or is missing.

OUTPUT

A comparison you can trace.

A report with references, honest uncertainty, and the working data behind it.

EXAMPLE PROMPT
Compare these documents against the criteria I supplied. Record where each factual claim comes from, distinguish source claims from your interpretation, and flag gaps. Save the comparison separately so we can revise it when I add another source.
Checking the result

03 / ACROSS YOUR DESKTOP

Prepare the file.
Check it in the app.

Create a report from your source files, open it in the app where it belongs, and refine the layout and content.

Operating other apps works on Windows. On macOS and Linux, the model can still prepare files and use the browser.

EXAMPLE WORKFLOW
03 / CONNECTED SURFACES

03 / CONNECTED SURFACES

FILESBROWSERAPPLICATIONS
INPUT

A file and an app.

Some sample files, the app or format you want, and clear limits on what may change.

METHOD

Prepare. Open. Review.

Build the output from the source material, then look at it in the app where it will be used.

OUTPUT

A finished file to review.

A new file, a summary of what changed, and an honest note about anything unfinished.

EXAMPLE PROMPT
Use these sample files to prepare a separate report in the format I specify. Open it in the supported application and check the layout and contents. Describe any action you could not verify. Keep the source files unchanged and ask before publishing or sending anything.
Staying in control of outside actions

04 / GOALS & AUTOMATIONS

Set it up once.
Let it keep going.

Turn a recurring job into an automation, or hand a bigger objective to a Goal and check in on its progress.

EXAMPLE WORKFLOW
04 / STANDING ORDERS

04 / STANDING ORDERS

SCHEDULERUNRECORD
INPUT

One clear instruction.

What to check or produce, where results go, and what must never change.

METHOD

Schedule. Test. Run.

Pick a schedule or trigger, run it once on demand to check it, then let it run.

OUTPUT

A record of every run.

Each run is logged, missed runs are flagged, and you can pause it at any time.

EXAMPLE INSTRUCTION · WEEKDAYS AT 08:00
Read any new files in my Inbox folder since the last run. Summarise each one in three bullet points and add the summaries to Reports/daily.md under today's date. Do not move, rename, or delete any files.
How automations work

Start small. Keep going.

Connect a model and try a short example first.

Getting started