Skill Deep-Dive 7 min read

The Poem That Could Have Been Written for Anyone

Generic AI verse isn't technically bad. It's contextless. The Poetry & Verse Prompt skill fixes that before the first line — by asking what it needs to know first.

SP
Founder, NovaKit
🎭
NovaKit Skill · Creative Bundle
Poetry & Verse Prompt — form-aware, occasion-calibrated verse that starts with three questions, not a blank page.
Quick answer: Generic AI verse isn't technically bad. It's contextless. The Poetry & Verse Prompt skill fixes that before the first line — by asking what it needs to know first.
In this guide

Poetry & Verse Prompt is a Claude AI skill — form-aware, occasion-calibrated verse that starts with three questions, not a blank page.

  1. What Generic AI Actually Produces — and Why
  2. What Form-Aware Generation Changes
  3. How the Skill Works in Practice
  4. Vanilla Claude vs Calibrated Output — Side by Side
  5. Is This Skill for You?
  6. Frequently Asked Questions

Your colleague is retiring after fourteen years. You've been asked to say something. You open Claude, type "write a retirement poem for a colleague who's been here fourteen years," and thirty seconds later you have eight lines about journeys and chapters, something rhyming "way" with "day," a mention of "legacy." It is not offensive. It is not wrong. It is the poem that could have been written for anyone who has ever retired, which means it was written for no one in particular — and your colleague will know that the moment they hear it.

That specific failure — technically competent, emotionally blank — is what uncalibrated AI verse produces every time. The problem isn't Claude's vocabulary or its grasp of metre. It's that it had no information about the actual person, the actual occasion, or the emotional note you were trying to hit. It defaulted to the statistical centre of what retirement poems look like across the training data, which is exactly as generic as that sounds.

The Poetry & Verse Prompt skill doesn't start by writing. It starts by asking what it needs to know.

What Generic AI Actually Produces — and Why

Uncalibrated verse has a recognisable signature. Abstract nouns — journey, light, path, chapter, legacy. Seasonal metaphors that apply to any transition. ABAB or ABCB rhyme schemes because they're the most statistically common. A sentiment that is broadly appropriate, which is another way of saying it fits every situation equally poorly.

The deeper problem is occasion-blindness. A wedding poem and a retirement poem and a poem for a child's first day of school are not variations on a single template — they're fundamentally different emotional tasks. A wedding poem is read aloud in front of everyone the couple loves; it needs weight, ceremony, something that earns the moment. A child's first-day poem needs to be brave and light and honest about the morning being hard. A retirement poem after fourteen years in the same building should know something about those fourteen years, or it's just applause in verse form.

💡
The core problem

Generic AI poetry fails not because Claude lacks vocabulary — it fails because it has no information about the specific person, occasion, or emotional register. Without context, every poem defaults to what works for everyone, which is what moves no one.

The tell, when you read generic AI verse, is that you could swap the recipient's name for anyone else's and the poem would still scan. That's the exact opposite of what makes a poem land.

That gap is exactly what the Poetry & Verse Prompt skill for Claude was built to close.

What Form-Aware Generation Changes

Poetry is a compression format. Every line carries more weight than prose because there are fewer of them — which means every structural choice matters. Whether a poem rhymes or doesn't. Whether it uses a fixed form or free verse. Whether the tone is dry and fond or quiet and tender. These aren't aesthetic preferences; they're decisions that determine whether the poem fits the room it's read in.

The Poetry & Verse Prompt skill researches current form conventions and what kinds of verse are resonating for different occasion types before it drafts anything. That's not pulling from cached training patterns — it's active research into what's actually landing: which wedding poem structures feel fresh versus tired, which imagery has been used so often it no longer carries weight, which approaches to grief are moving people right now versus which ones feel formulaic.

The difference between a poem that makes someone cry and one they politely thank you for is almost never the vocabulary — it's whether the form and register were chosen for them.

Then come the calibration questions. Before a single line is drafted, the skill asks three things: who this is for, what the occasion is, and what emotional register it should hit. Warm and funny, or quiet and reflective? Should it rhyme — or would rhyme make it feel forced for this particular person? Those answers route the skill to an entirely different output than any other combination would produce. The same skill writes completely differently for a retirement poem read at a pub lunch versus one read at a formal dinner.

How the Skill Works in Practice

The skill is built around one principle: earn the right to write by understanding the situation first. Here's the sequence.

1
Three-question calibration
Who is this for, what is the occasion, and what emotional note should it hit — tender, funny, formal, intimate, bittersweet? These three inputs route the skill away from the generic centre and toward something specific. Two different sets of answers produce two poems that share almost nothing in common.
2
Live form research
The skill researches which forms are landing for this occasion type right now — not defaulting to ABAB because it's statistically common, but making an active choice about what structure will serve this specific moment. Overused imagery gets flagged; approaches that feel current get prioritised.
3
Multiple form variants
The skill produces two to three versions across different forms — typically a free-verse version that prioritises emotional specificity alongside a structured version that uses form to create ceremony. Each comes with a brief note on why it makes the choices it does, so you can choose with understanding rather than guesswork.
4
The quality gate
Before output reaches you, an automatic quality check runs against the calibration brief. Forced rhymes, weak imagery, lines that don't earn their place — rewritten before delivery. The version you receive already passed; you're not starting from a draft that needs rescue.
NovaKit Skill · Creative Bundle
Poetry & Verse Prompt for Claude
Three calibration questions. Multiple form variants with structural notes. Works inside your existing Claude account — no setup, no subscription.
See the skill $5 · instant download

Vanilla Claude vs Calibrated Output — Side by Side

The same brief: a poem for a colleague leaving after fourteen years, read at their leaving lunch — warm, a little dry, not too formal. Here's the gap.

Without the Skill
Fourteen years have passed like fleeting days,
A journey filled with hard work and with praise.
You've walked the path with grace and with care,
And left your mark beyond compare.

As you move on to chapters that are new,
We'll cherish every memory of you.
Though roads diverge and time moves ever on,
Your legacy remains though you are gone.
Correct metre. Working rhymes. Could describe anyone who has ever left any job anywhere.
✓ With the Skill
Fourteen years of knowing where the good mugs are,
which meetings end early, whose birthday cake
is worth staying late for.

You became the kind of colleague
people mean when they say they miss the place —
not the place, exactly.

The next room gets the version of you
we made together.
Take that with you. It's not small.
Occasion-specific, register-calibrated. The "good mugs" line signals this was written about someone real.

The left version is technically fine. The rhymes land, the sentiment is appropriate, the form is conventional. But there's nothing in it that could only have been written about this person, this workplace, this leaving. The right version earns its emotional landing — "the next room gets the version of you we made together" — through specificity, not sentiment. That shift comes entirely from the calibration questions answered before a word was drafted.

Notice also what the calibrated version doesn't do: it doesn't force a rhyme where one would feel strained. The brief said "a little dry" — and the form reflects that. Free verse with conversational line breaks reads dryer than ABAB. The skill chose the form because the brief called for it, not because it was the default.


Is This Skill for You?

The Poetry & Verse Prompt works best for people who care about getting it right but don't write verse habitually. It's not a replacement for a poet's judgment — it's the research and structure layer that gets you from "I have no idea where to start" to "I have two strong versions and I understand the trade-offs between them."

This is for you if
  • You've been asked to write something for a specific occasion and the stakes feel real
  • You've tried vanilla Claude and got output that felt generic or forced
  • You want to understand why certain forms work for certain occasions, not just receive an output
  • You write verse occasionally and want a structured starting point before doing the final craft work
This probably isn't for you if
  • You want a poem generated in one click with no questions asked
  • The occasion is low-stakes and any verse will do
  • You're a working poet who already has strong form instincts

The calibration table below shows how the same skill produces fundamentally different outputs depending on the inputs — which is the point.

Occasion Register → Form choice What the skill prioritises
Wedding reading Tender, formal → structured stanza, restrained rhyme Ceremony; lines that earn the moment they're read in
Colleague retirement Warm, dry → free verse with conversational line breaks Specificity; something that couldn't describe anyone else
Child's birthday Playful, light → rhythmic, singable, short stanzas Energy and delight; form matches the reader's age
Eulogy verse Quiet, reflective → unrhymed, unhurried, space between lines Restraint; let grief breathe rather than resolve it
Brand or launch copy Sharp, contemporary → prose-poem rhythm, no forced rhyme Voice distinctiveness; verse that doesn't read as verse

Frequently Asked Questions

Can AI actually write good poetry?
AI can write technically competent verse — correct metre, working rhyme schemes, appropriate sentiment. The limitation is context: without knowing the specific person, occasion, and emotional register, output defaults to universal imagery that fits everyone and therefore moves no one. The skill closes that gap by gathering context before generating.
Why does AI poetry sound so generic?
Generic verse pulls from the statistical centre of what poetry looks like across training data — ABAB rhyme, seasonal metaphors, abstract nouns like "journey" and "legacy." It's calibrated for the average occasion, which means it's wrong for any specific one. The fix is context-first generation, not better vocabulary.
Do I need to know anything about poetry forms to use this skill?
No. The skill handles form selection based on your calibration answers. You answer three questions about person, occasion, and register — the skill decides whether free verse, a sonnet structure, or a conversational prose-poem serves your brief best, and explains why in the output notes.
What does the skill actually produce?
Two to three verse variants across different forms, each with a brief structural note explaining what the form communicates and why it was chosen for your brief. You see the options, understand the trade-offs, and choose — or combine elements from more than one.
Does this work for professional or commercial verse — not just personal occasions?
Yes. The calibration handles brand voice, launch copy, editorial verse, and anything else where tone and form need to be deliberate. The "occasion" input is flexible — it covers formal personal occasions, creative projects, and commercial contexts equally.

The best poems at weddings and funerals and leaving lunches make people go quiet because they feel like they were written for that room, that person, that exact morning. That's not something that comes from a better vocabulary or a longer prompt. It comes from knowing the context before the first line — and then building every choice from there. That's what the calibration step is for. The writing follows.

The next piece most people tackle from here is a children's book with the rhythm young readers follow.

Ready to try it?
Poetry & Verse Prompt for Claude
Multiple form variants, structural notes on each, quality gate built in. Works with your existing free or Pro Claude account. Done in minutes.
Get the skill $5 · instant download · 7-day refund

Put this to work: the Poetry & Verse Prompt skill for Claude turns everything above into one guided workflow you run in a normal Claude chat. Not ready to buy? Start with a free Claude skill and see how it works first.

Tags Poetry Writing Claude AI Creative Writing Verse & Form AI Skills
Free skill
Try NovaKit before
you spend a dollar.

Get the LinkedIn Post Engine free — the same skill that runs live trend research before every post. Drop your email and it lands in your inbox in seconds.

💼
LinkedIn Post Engine
Social · normally $9 · free today
Live trend research before every post
Hook variants calibrated to what's converting this week
Works on a free Claude account

No spam. No account. Unsubscribe any time.