How to Write a Product Description with ChatGPT: A Workflow

What You Will Deliver

This workflow turns a bare product into a description that actually sells. By the end you will have a fact sheet, a benefit-led draft, an SEO-ready version, and platform variants (Amazon, Shopify, social) you can paste anywhere. The whole thing takes 15 to 30 minutes per product and you can repeat it for every SKU in your catalog.

You only need a ChatGPT account. The trick is not to ask it to “write a product description” in one shot, which gives generic fluff. Instead, you feed the model one job at a time: facts first, then angle, then draft, then SEO, then variants. For deeper control over voice and length when drafting longer copy, the best ChatGPT settings for long-form writing guide is useful background.

Step 1: Build the Product Fact Sheet

Before any selling language, capture the raw facts. A model cannot invent accurate specs, so you must supply them. This step produces a clean brief you will reuse in every later prompt.

You are a senior e-commerce copywriter. I will give you raw product facts. Turn them into a structured fact sheet.
Product: [NAME]
Category: [CATEGORY]
Key specs: [LIST MATERIALS, SIZE, WEIGHT, COLORS, ETC.]
Top 3 features: [FEATURE 1], [FEATURE 2], [FEATURE 3]
Who it is for: [TARGET BUYER]
Problem it solves: [ONE SENTENCE]
Price band: [BUDGET / MID / PREMIUM]
Return the fact sheet as a compact list with these headings: Specs, Features, Buyer, Problem solved, Positioning.

You get a tidy brief. Save it. If the model adds claims you did not give it (like “award-winning” or “best on the market”), delete them — those are hallucinations you do not want in a live listing.

Step 2: Choose the Angle (Features vs Benefits)

Decide whether the description leads with what the product is or what it does for the buyer. Benefit-led copy converts better for emotional purchases; spec-led copy wins for technical buyers. Pick one so the draft stays consistent.

Using this fact sheet: [PASTE FACT SHEET]
Propose 3 description angles:
1. Benefit-led (focus on the outcome for the buyer).
2. Spec-led (focus on materials, build, technical proof).
3. Story-led (place the product in a real moment of use).
For each angle, write one sentence that could open the description and name the reader emotion it targets.
Then recommend which angle fits a [BUDGET / PREMIUM] buyer and why.

Now you have a clear opener direction. Choose the angle that matches your price band. A premium skincare item usually wants benefit or story; an SSD drive wants spec-led proof.

Step 3: Draft the Description

With the angle locked, generate the full body. Keep it scannable: short paragraphs, a benefit list, and one quiet CTA. The ChatGPT rewrite vs edit guide explains why vague “make it better” instructions fail — here we give exact rules instead.

Write a product description using this fact sheet: [PASTE FACT SHEET]
Angle: [CHOSEN ANGLE FROM STEP 2]
Rules:
- 120 to 180 words.
- Open with the chosen angle sentence.
- Follow with 4 to 6 short bullet benefits (one line each, start with a verb).
- Then 1 paragraph of proof (specs that back the benefits).
- Tone: confident, plain English, no hype words like "amazing" or "revolutionary".
- End with one soft call to action (e.g. "Add to cart and see the difference").
Return clean HTML: h2 for the title, p for paragraphs, ul for bullets.

You now hold a solid first draft. Copy it to your editor. Expect to trim repetition — models love restating the benefit. The draft is a base, not a final publish.

Step 4: SEO-Optimize the Copy

Marketplaces and Google reward descriptions that use the search phrase naturally. Add the keyword without stuffing, and write a meta line for search snippets.

Take this description: [PASTE DRAFT]
Target keyword: [KEYWORD, e.g. "wireless noise cancelling earbuds"]
Tasks:
1. Weave the keyword into the opening sentence and one bullet, naturally.
2. Do NOT exceed a 2% keyword density.
3. Write a meta description under 155 characters that includes the keyword.
4. Suggest 3 secondary long-tail keywords to mention if I expand the page.
Return: optimized HTML description, then a separate line for meta description, then the 3 keywords.

You get a search-friendly version plus a meta line. Keep the keyword in the title tag on your store too — that is where it matters most.

Step 5: Generate Platform Variants

One description rarely fits every channel. Produce a short Amazon-style version, a Shopify version, and a social caption. The AI marketing tone and persona guide helps you tune voice per channel if you want more control.

From this description: [PASTE OPTIMIZED DESCRIPTION]
Create 3 variants:
A. Amazon version: 200 to 300 words, includes a "Why buy" subhead, bullet specs first.
B. Shopify version: 100 to 150 words, brand-voice, one CTA button phrase.
C. Instagram caption: under 80 words, 3 hashtags, one question hook.
Label each variant clearly. Keep facts identical across all three.

You now have channel-ready copy. Paste each into its destination. Always keep the facts consistent — different wording is fine, different specs are not.

Final Deliverable Checklist

Common Pitfalls

1. Skipping the fact sheet. Without hard specs the model invents them, and fake claims get listings removed.

2. Asking for the whole listing in one prompt. You get a shallow template. Stage it: facts, angle, draft, SEO, variants.

3. Keyword stuffing. Forcing the phrase five times reads as spam and hurts ranking. Two natural mentions win.

4. Copying the draft verbatim. Always fact-check specs and trim filler before publishing.

Sources & Notes

Prompts tested with ChatGPT (GPT-4o and GPT-4.1), free and Plus tiers, in October 2026. The workflow follows a five-stage delivery pattern (facts, angle, draft, SEO, variants) and is model-agnostic — the same prompts work in Claude or Gemini. Category mapping follows the guunn.com AI Prompts plan (Writing = 956). No live APIs were called during article creation; all prompts are copy-paste text for the reader.