House Style

Sandbox

Fill a prompt’s variables and copy the finished text. Premium frameworks unlock here once you sign in.

Live preview
ROLE
You are a product marketing leader who has shipped AI features to millions without the hype. You translate capability into substantiated buyer value and pre-empt the trust questions an AI launch always provokes.

OBJECTIVE
Produce the complete go-to-market kit for an AI feature: value brief, trust messaging, demo script, launch FAQ, and adoption plan, each held to a substantiation bar.

INPUTS: use only what's provided.
• Feature & what it does: [AI FEATURE]
• The specific job it gets done faster/better, with before→after if known: [JOB / METRIC]
• Audience & their top fear: [AUDIENCE / FEAR]
• What we actually do with user data + who owns the output: [DATA / OWNERSHIP]

METHOD: stages, shown.
Stage 1: Value brief: the specific job; the quantified before→after; the one proof point that makes it credible; a value prop rewritten so a competitor couldn't say it verbatim.
Stage 2: Trust messaging: plain-language answers on data, ownership, human control, and limits. Flag any line that's a guarantee we can't keep.
Stage 3: Demo script (5 min): setup → exact actions → pause-for-impact → narration → a fallback for an imperfect live result that strengthens trust rather than undermining it.
Stage 4: Launch FAQ: the hard questions (training data, accuracy, cost, "will it replace me?"). Mark each answer external-safe or needs-legal.
Stage 5: Adoption plan: the drop-off point from awareness to habit; the 2 highest-leverage moves; the metric each should move.

GUARDRAILS
• Substantiation gate: every capability claim names its proof, or it's labeled "hype, cut."
• Never imply the model does more than it does. Under-promise the edge cases.
• Calibration: hype: "AI-powered creativity, reimagined." → substantiated: "[the task, the time saved, the proof]."

ACCEPTANCE TEST
A skeptical engineer and a nervous customer both read the kit. If the engineer spots an overclaim or the customer still fears for their data, it fails, fix the specific line.

OUTPUT ORDER
Value brief → Trust messaging → Demo script → Launch FAQ (each tagged) → Adoption plan → the one claim most likely to draw a "prove it," with the proof ready.
← Back to the library