Using Fanzy AI to draft product descriptions you then cut down

The product is finished. The file is uploaded, the price is set, the cover looks fine. Then there is a description field with a blinking cursor in it, and it is a quarter past eleven, and you write four sentences, delete three, and close the laptop. The product does not go live that night, or that week. Almost nobody admits that a text box is what stopped them, because it sounds ridiculous, and it is nonetheless the most common reason a finished product sits unpublished.

Why the description field is harder than it looks

Writing about your own work requires you to hold two incompatible positions at once. You know the thing intimately, which makes everything about it feel obvious and therefore not worth saying. You are also trying to persuade, which makes you reach for language you would never use in a message to a friend. The result is either three flat lines that describe a file, or a paragraph of confident nothing.

The honestly difficult part is that a good description contains information you have not written down anywhere yet: who this is not for, what it does not include, what state the buyer is in when they need it. You cannot look that up. It comes out of you, usually in the middle of answering a message, and then it is lost in a chat thread.

So the useful division of labour is this: let the assistant handle structure and momentum, and keep the specifics for yourself. Fanzy drafts copy as well as designing the storefront, which means the blank field stops being a blank field. What it cannot do is remember the thing you told someone in a DM last Tuesday.

From scratch, or from a draft

Both routes end with you writing. The difference is where the effort lands: inventing sentences from nothing, or reacting to sentences that are already there. Reacting is faster and, oddly, produces braver copy, because it is easy to be ruthless with someone else's words.

The workflow, step by step

Do this one product at a time. It takes about ten minutes per product once you have done it twice, and the tenth minute is the one that matters.

1. 1Give Fanzy the object and the buyer, not the vibe. "A 42-page PDF of daily mobility routines, ten minutes each, for people who sit at a desk and have started noticing their hips." That single sentence carries more than a paragraph of adjectives.
2. 2Ask for it longer than you need. A draft that overshoots gives you material. A draft that is already tight leaves you nothing to choose between.
3. 3Paste it in and read it aloud once, without editing. You are listening for the sentence that makes you wince and the one that is accidentally true. Mark both.
4. 4Run the four cuts below. This is the actual work and it takes about six minutes.
5. 5Check it against the card, not the editor. Storefront and product pages are statically rendered with their own titles and descriptions, so open the live page on a phone and see how much of your text survives above the fold.
6. 6Publish it slightly imperfect. A live product with adequate words earns you information; a perfect draft in a tab earns you nothing.

A worked example: a template pack at 24

Suppose you sell a set of client-onboarding templates for freelance designers. The draft comes back with a smooth opening about streamlining your workflow, a middle paragraph about professionalism, and a close about saving time. Nothing is wrong with it. Nothing in it would make a designer reach for a card either, because every sentence would be equally true of a competitor's pack.

Now the four cuts. The opening goes. The claims become "nine documents: proposal, contract, kick-off questionnaire, three-round revision policy, handover checklist". The buyer's moment goes in near the top: "for the project where the client asked for a fourth round and you had nothing written down to point at". The close states delivery plainly. The description drops from 180 words to about 90, and it now says something only someone who has been through that argument would think to say. If ten people a week look at the card and one buys, you have per-product analytics telling you which half of that is the copy's fault — views, buyers and revenue are broken out per product on the analytics page.

The mistake most people make

The mistake is batching. You ask for all six descriptions at once, they come back in one voice with one rhythm, you tidy them lightly, and now your storefront reads like a catalogue printed by a company rather than a page made by a person. Every card opens the same way. Every card has the same three-beat close. Visitors do not consciously notice; they just stop reading after the second card. Draft one product, cut it, publish it, then come back tomorrow for the next. The unevenness between them is not a flaw — it is the main thing that signals a human made the page.

Treat the draft as scaffolding. It exists so that the hard part — deciding what is true and specific about your product — happens with something on the screen instead of nothing. The four cuts take six minutes and they are the same six minutes every time, which is what makes this repeatable rather than a one-off burst of motivation. Which plan you are on changes none of it, and 0% platform fee applies across all of them — the pricing page has the rest.