SSkilvy

There are many ideas — the point is to pick a living one

AI can produce a hundred ideas in a minute. But not all are equal. A strong premise rests on three things: an interesting hero, a clear conflict and meaningful stakes (what the hero loses or gains). Set these frames — and the ideas get deeper.

Genre drama, fantasy, mystery Hero who is at the center Conflict what stands in the way Stakes what they lose or gain
A good premise is a hero, a conflict and stakes. AI produces options in batches.

Ask not for "an idea" but for a construction

Give 4 premises for a short story: genre drama, theme second chance. For each, state the hero, conflict and stakes.
1) A pianist has lost her hearing but gets a chance to play a last concert (conflict: body vs dream; stakes: farewell to her former self). 2) A father finds a letter from his dead son with an unread request (conflict: guilt vs time; stakes: peace or a lifelong burden). 3) A former thief returns to a town that remembers him (conflict: past vs new life; stakes: acceptance or exile). 4) A teacher recognizes a former student in a homeless man (conflict: walk past or step in; stakes: guilt or rescue).

Develop the chosen idea with questions

Once you pick a premise, do not ask "write the story". Deepen it in parts: "what if the hero is to blame for the conflict?", "what surprising twist is possible?", "what does the hero realize at the end?". Dialogue creates depth.

Look for the "what if"

The best stories often start with an unusual question. Ask AI: "give 10 surprising what-ifs on this theme". Many will be filler, but one or two will spark — and that is already your unique story.

Do not take the first idea. Generate 10, pick the one that resonates with you personally — readers feel when the author truly cares.

🧠 What does a strong premise rest on?