You've captured a thought and decided it's worth developing. Pasting it into a free AI model is the fastest way to turn one sentence into a page of thinking — but most people get mush back, because they paste two words and ask "thoughts?"
The model is only as good as what you give it. Here's the anatomy of a prompt that works, then four you can lift directly.
The three ingredients of a good prompt
- A role. Tell the model who to be — "act as a product strategist," "act as a sceptical investor." This narrows its enormous range down to the voice you actually want.
- Your raw material. Paste the real, unedited note. This is where voice capture wins: your spoken note is full of context — the why, the use case, the offhand aside — and the model thrives on exactly that detail.
- A defined output. Ask for a specific shape: "a 3-step roadmap," "exactly 5 competitors," "the 3 biggest risks." Vague asks get vague answers; specific asks get usable ones.
Never trim your own note before pasting it. "Wine app" gives the model nothing. The full dictation — "an app that reminds you which wines you liked because I forget the good ones by reorder time" — gives it a problem, a user and a motivation. Paste it whole.
Prompt 1 — The roadmap
The workhorse. Turns a raw note into a plan, competitors and challenges in one shot.
I captured this raw thought while driving: "[paste your full Note Now dictation here]" Act as a product strategist. Turn this into a 3-step actionable roadmap, identify 3 likely competitors, and list the 3 immediate challenges I'd need to solve first. Keep it concise and flag any assumptions you're making.
Prompt 2 — The competitor scan
When you want to know what already exists before you fall in love with the idea.
Here's a rough idea I had: "[paste note]" List 5 existing products or services that already solve some version of this. For each, give one line on what they do well and one line on the gap they leave open. Then tell me, in one sentence, where my idea could realistically be different.
Prompt 3 — The reality check
The one most people skip and most need. Ask the model to argue against you.
Idea: "[paste note]" Act as a sceptical investor who wants me to succeed but won't flatter me. What are the 3 strongest reasons this fails or isn't worth doing? For each, suggest the cheapest way I could test whether that risk is real this week.
Prompt 4 — The first concrete step
For when the idea feels real but you're frozen on what to do Monday morning.
Idea: "[paste note]" I have about 2 hours and no budget. Give me the single most valuable thing I could do in those 2 hours to move this forward — not research, an actual output I can point to afterwards. Be specific.
Treat the output as a draft, not an oracle
Everything that comes back is a strong first draft and nothing more. Models invent confident nonsense, miss the obvious, and occasionally flatter you. Read it the way you'd read a sharp colleague's quick take: useful, worth arguing with, not the final word. Keep what's true, delete what isn't, and the page in front of you is now genuinely yours.
This expand step slots straight into the wider free workflow — see turning a CarPlay note into a business proposal for where it sits in the chain, and free AI prototyping for taking the output further into something clickable.
The richer the raw note, the better the AI gets.
Because Note Now captures your exact spoken words, your raw text is full of the context AI models thrive on. Capture it hands-free, paste it whole. Free to download.
Download Note Now — Free