Keep the evidence close.
Make it possible to return to a source, inspect a claim and see what remains uncertain.
A little of the
thinking behind it.
Skills, tools and experiments in making AI useful. Keep the thinking visible, make the thing with care, and leave someone a way in.
Good work has
a next day.
A first draft gives you something to react to. The next questions matter too: where the evidence came from, why a choice was made, what was checked, and how someone else can continue.
I build capabilities around those moments. Some help with research and judgment. Some help build software or direct creative work. Others keep the evidence, decisions and next steps close to the project.
They have different jobs and different shapes. You can start with one that fits the work in front of you.
Choose a shape that fits.
A useful extension can be a small skill, a tool or a shared library. Let the job decide.
Improve with a reason.
Make the change intentional. Explain what it should help, and what would show that it did.
Make the next step clear.
A person returning tomorrow should have a place to begin, with the current state and the important decisions visible.