Learn what the award winners in your niche actually share

Reverse a folder of the best work in your category, keep it as a corpus, then ask what the winners do that everyone else does not.

A shortlist is forty images you admire and cannot act on. Reversed into a corpus you can question, it becomes the brief for your next campaign, and it keeps paying out every time you open it.

Why this will not look generic: The brief is synthesised from work that a jury of practitioners already judged to be the best in your category, so the specification comes from demonstrated craft rather than from the model's average of everything.

What you walk away with

One prompt per award-winning piece, kept as a folder in your library
The craft choices the winners share, which ordinary work in the category does not
A synthesised brief you can generate from
A corpus that keeps answering new questions later

Typical run: 20 to 60 pieces. Reverse in bulk over /v1 first, then do the reasoning in a thread; forty paragraphs is a comfortable read for one conversation.

This one happens in one conversation

You bring: A folder of the best work in your category, saved to your library.

Open with: Read every prompt in my 'award winners' folder. What craft choices show up again and again, and which of them would a competent but unremarkable version of this work get wrong?

What the thread holds on to:

All forty prompts at once, which is the only way a pattern across them is visible
The distinction it drew between what is common and what is actually good
Your own brand directive, so the synthesis comes out in your voice rather than theirs

Then you can just ask:

Which of these choices are specific to the category and which are just good craft?
Write me a brief that uses four of them on my product
Generate three options from that brief
Now which of these would still work at thumbnail size?

What the reverse actually gives you

Point it at: A folder of the shortlisted work in your category, reversed one by one.

What comes back (representative example): Across the set, the recurring choices are not the obvious ones. Thirty-one of forty place the product off-centre and let negative space carry the composition. Twenty-eight use a single hard light source with the shadow left intact rather than filled. Twenty-four limit themselves to two colours plus a neutral. Only six use the soft, evenly-lit, centred treatment that dominates ordinary category work. The restraint is the pattern: fewer elements, harder light, more empty space, and one idea per frame rather than three competing for attention.

Things you would never have thought to type:

shadows left intact rather than filled, in 28 of 40
negative space doing the compositional work, not the subject
two colours plus a neutral as a discipline, not a palette
one idea per frame, where ordinary work puts three

What you change: Your product and your idea. The restraint is the finding, and restraint is the hardest thing to specify from scratch because it is defined by what you leave out.

How it works

Collect the shortlist (In Chrome) - Right-click through the winners as you read them, which is when you actually have the tab open.
File them as a corpus - Saving is session-authed web only. Put them in one folder and tag it, because the folder is the thing you will query later.
Ask the set what it knows (In Claude) - The reasoning is the deliverable here, not any single prompt.

Why not just keep a moodboard?

A moodboard is forty things you liked. It cannot tell you why you liked them, and it certainly cannot tell you what thirty-one of them quietly have in common.

A corpus answers questions a folder cannot. Once the set is reversed, the collection is comparable text rather than forty impressions. You can count how many use hard light, ask which choices are category convention versus genuine craft, and get a number back instead of an opinion.
It gets more valuable the longer you keep it. This is the one thing here that compounds. A folder you add to for a year becomes a reference you can interrogate from any surface, and the questions you ask of it in month twelve are not the ones you asked in month one.
Synthesis, not imitation. Copying one winner gets you a derivative of that winner. Reading forty and pulling out the shared craft gets you a brief, which is a different and much more useful object, because it applies to your product rather than reproducing theirs.

Worth doing once a year with the current shortlists, and worth keeping forever once you have it.

Cost and time

This workflow runs at from 5 credits + 2 reverses + about 1 minute. Your result carries automatically from each step into the next, so there is no re-uploading between tools. Reverses are billed from your reverse meter and generation steps from your AI Credits, both bundled in every pay-as-you-go credit pack.

Browse all guided workflows or all individual tools. Pay as you go, no subscription, AI Credits never expire.