7 Best Claude Prompts for Ecommerce Product Descriptions

I spent 47 minutes last Tuesday writing a single property description. One listing. A three-bedroom villa in Câmara de Lobos with a sea view and a renovation history that required careful language for both Portuguese and international buyers. That kind of copy used to eat my mornings. Then I started building a specific prompt library inside Claude, and that same type of description now takes me under 8 minutes. If you sell products online, your situation is not so different — you are writing variations of the same structured text, over and over, for dozens or hundreds of SKUs. The right Claude prompts cut that time dramatically. Here is exactly what works, based on hands-on testing through early 2026.

Why Claude Outperforms Other AI for Product Copy

I have used ChatGPT, Gemini, and Claude side by side for writing-heavy tasks. For product descriptions specifically, Claude consistently produces copy that sounds less formulaic. It handles nuance better — things like tone shifts between a luxury item and a budget one, or the difference between writing for a wholesale buyer and a retail consumer.

Claude’s training leans toward careful, precise language. That matters a lot when you are writing product copy where a wrong word creates a return or a complaint. It also follows complex multi-part prompts more reliably than most models I have tested. You can give it a structure with six specific requirements and it will follow all six, not just the first three.

The model I use most is Claude 3.5 Sonnet via the claude.ai interface. As of early 2026, it sits at $20/month for the Pro plan. For the volume of output it produces in my business, that cost pays for itself within the first two or three working days of any given month.

The Anatomy of a High-Converting Product Description Prompt

Most people write weak prompts. They type “write a product description for this water bottle” and wonder why the output sounds generic. The prompt is the problem. A well-built product description prompt has five components:

1. Role assignment — Tell Claude who it is writing as
2. Audience definition — Who is reading this and what do they care about
3. Product data — Specs, materials, dimensions, key differentiators
4. Format requirements — Word count, structure (headline, bullet points, paragraph), tone
5. Platform context — Shopify, Amazon, Etsy, WooCommerce all have different requirements

Leave any of those out and the output degrades. Add all five and you get something usable on the first pass, 80% of the time in my experience.

7 Best Claude Prompts for Ecommerce Product Descriptions

1. The Shopify Standard Listing Prompt

This is my workhorse. I adapted it from a real estate listing format I built for myself, and it translates directly to product copy.

You are an experienced ecommerce copywriter specializing in Shopify product pages. 

Write a product description for the following item:

Product name: [NAME]
Category: [CATEGORY]
Key specs: [LIST SPECS]
Main benefit: [ONE SENTENCE]
Target buyer: [WHO THEY ARE AND WHY THEY SHOP HERE]
Tone: [e.g., professional and approachable / playful and casual / premium and understated]

Format:
- Opening sentence that leads with the main benefit (not the product name)
- 3-5 bullet points highlighting features with brief explanations of why they matter
- 1 closing paragraph (2-3 sentences) that reinforces the purchase decision
- Total: 150-200 words

Do not use the words "perfect," "amazing," or "elevate." Avoid exclamation marks.

The instruction to avoid specific filler words is critical. Claude will use “perfect” and “amazing” constantly if you do not block them. Adding that line to every prompt saves you editing time on the back end.

2. The Amazon SEO-Optimized Bullet Points Prompt

Amazon listings live and die by their bullet points. This prompt generates five bullets structured around Amazon’s character limits and keyword placement conventions.

You are an Amazon marketplace copywriter who understands SEO ranking factors.

Write 5 bullet points for an Amazon product listing. Each bullet must:
- Start with a capitalized feature name in 2-4 words (e.g., "LEAK-PROOF DESIGN —")
- Be 150-200 characters long
- Lead with a benefit, then support it with the feature or spec
- Include one of these keywords naturally in at least 3 of the 5 bullets: [INSERT TARGET KEYWORDS]

Product: [NAME AND SPECS]
Target customer: [DESCRIBE]
Main use case: [HOW THEY USE IT]

I have run this prompt for a client who sells handmade ceramic pieces through Amazon Handmade. She produced 12 new listings in one afternoon using this format. Her average bullet-writing time dropped from about 25 minutes per listing to under 6 minutes.

3. The Etsy Story-First Description Prompt

Etsy buyers respond to story and context more than technical specs. This prompt is built for handmade, vintage, or artisan products where emotion drives the purchase.

You are a copywriter who specializes in Etsy product listings for handmade and artisan sellers.

Write a product description that opens with a 2-3 sentence story or scene — not a feature list. The reader should feel something before they read any specifications.

After the opening, include:
- What the item is made of and how
- Dimensions and practical details
- Who it makes a great gift for (if applicable)
- Care or usage instructions in plain language

Tone: warm, personal, like it was written by the maker themselves
Length: 200-250 words

Product details: [INSERT ALL RELEVANT INFORMATION]

4. The Bulk Variation Prompt for Multiple SKUs

This is where Claude’s ability to follow complex structured instructions really pays off. If you have a product in 8 colors or 5 sizes, you do not want 8 identical descriptions with just the color word swapped out. That looks thin to Google and lazy to buyers.

I need product descriptions for [NUMBER] variations of the same product. Each description must be unique — do not simply swap the variable word. Instead, highlight a different use case, mood, or customer profile for each variation.

Base product: [CORE PRODUCT DETAILS]
Variations: [LIST EACH VARIANT WITH ANY UNIQUE SPECS OR ATTRIBUTES]

For each variation, write:
- A unique opening sentence (no two can start the same way)
- 3 bullet points specific to that variant where possible
- 1 sentence call to action

Format each one clearly with the variation name as a header.

5. The Technical Product Prompt for B2B Buyers

Selling tools, equipment, industrial supplies, or anything where the buyer knows exactly what they need? This prompt switches the register from persuasion to precision.

You are a technical writer for a B2B ecommerce catalog. Your reader is a procurement professional or experienced practitioner who does not need to be sold to — they need accurate, complete information to make a buying decision.

Write a product description that prioritizes:
1. Exact specifications (format as a structured list)
2. Compatibility and use-case parameters
3. What this product is NOT suitable for (this builds trust with technical buyers)
4. Standards, certifications, or compliance details if provided

Tone: precise, neutral, no marketing language
Format: short intro (1-2 sentences max), then structured specs, then 1 paragraph on applications

Product data: [INSERT]

The instruction to include what the product is NOT suitable for is something I borrowed from how I write property listings for commercial buyers in Madeira. Buyers who know their field trust you more when you acknowledge limitations upfront. It works exactly the same way in B2B product copy.

6. The SEO Meta Description Companion Prompt

After you generate the main description, run this immediately after in the same Claude conversation. It uses the context already established.

Using the product description you just wrote, create:

1. An SEO meta description (150-160 characters, include the primary keyword "[INSERT KEYWORD]", end with a soft call to action)
2. An SEO page title (50-60 characters, keyword near the front)
3. Three alt text options for the main product image (each under 125 characters, descriptive and keyword-relevant)

Do not start the meta description with the product name.

7. The Rewrite-and-Improve Existing Copy Prompt

Most ecommerce stores already have descriptions — they are just bad. This prompt is for improving what exists rather than starting from scratch.

Here is an existing product description that needs improvement:

[PASTE EXISTING DESCRIPTION]

Problems to fix:
- [e.g., too vague / reads like a spec sheet / uses passive voice throughout / no clear benefit statement]

Keep:
- [e.g., the technical specifications / the brand tone / any claims about certifications]

Rewrite it with:
- An active voice throughout
- Benefits leading each section before features
- Word count between [X] and [Y] words
- The same overall information, better organized and more compelling

Do not invent any specifications or claims not present in the original.

That last line matters. Claude will sometimes embellish. Telling it explicitly not to invent information keeps your output accurate and avoids a liability problem if specs get fabricated.

Prompt Comparison: Which Format Works Best for Each Platform

Platform	Best Prompt Type	Key Format Requirement	Avg. Time with Claude
Shopify	Prompt #1 (Standard Listing)	Benefit-first opening, scannable bullets	5–8 min per listing
Amazon	Prompt #2 (SEO Bullets)	Capitalized feature headers, 150-200 chars	4–6 min per listing
Etsy	Prompt #3 (Story-First)	Emotional opening, maker voice	6–10 min per listing
WooCommerce/B2B	Prompt #5 (Technical)	Specs-first, neutral tone, limitations included	5–7 min per listing
Any platform (bulk)	Prompt #4 (Variations)	Unique angle per variant, no swapped-word clones	15–25 min for 6–8 variants

My Real-World Experience Using These Prompts

I want to be specific here because I see too many “AI productivity” articles that claim massive time savings without any real context behind the numbers.

In February 2026, I took on a project outside my core real estate work — a friend who runs a small online shop selling locally-made Madeiran products, mostly embroidery, wickerwork, and artisan food items, asked me to help her rebuild her Shopify product pages. She had 34 listings. Most of them were essentially a paragraph copied from a supplier sheet or, in some cases, just a sentence and a price. Not good for search visibility, not good for conversion.

I had one Saturday to work on this with her. I built out a prompt template using a version of Prompt #1 above, customized for her brand — Portuguese artisan products aimed at tourists, expats, and international gift buyers. I added a brand voice section to the prompt: “warm, specific, rooted in Madeiran culture, avoid generic phrases like ‘one of a kind’ or ‘traditional craftsmanship.'”

We worked through the listings together. She would read me the product details, I would paste them into the prompt, Claude would generate the description, and she would read it aloud and flag anything that felt off. That review loop took about 2 minutes per listing on average. Some needed minor edits — a word that did not sound like her, a spec that was slightly wrong. A few came out near-perfect on the first pass.

In 4.5 hours, we had completed all 34 listings. Each one had a proper benefit-led opening, a set of scannable bullets, and a closing sentence. We also ran the SEO meta description companion prompt (Prompt #6) for each one. Total output: 34 product descriptions, 34 meta descriptions, and 34 page titles.

Before this session, she had estimated it would take her 3 full days to do this alone, and she had been postponing it for two months. Her previous approach was writing each one manually, which took her 45-60 minutes per listing when she factored in looking things up, getting stuck on phrasing, and second-guessing herself.

At her rate of roughly 45 minutes per listing, 34 listings would have been about 25 hours of work. We did it in 4.5 hours between two people, with Claude doing the heavy lifting on the draft. That is a real, measurable time difference — not an estimate inflated for impact.

The direct translation to my own real estate business has been similar. I write property descriptions in two languages (English and Portuguese) for every listing I take on. Claude handles the English draft from my structured prompt, and I write the Portuguese version myself because machine translation still does not capture the local real estate register properly. That alone saves me between 30 and 45 minutes per listing on the English copy, which adds up to about 5-6 hours a month given my typical listing volume.

The Real Limitations I Hit When Using Claude for Product Copy

I will not pretend these prompts are magic. There are three genuine limitations I have run into repeatedly.

Claude does not know your brand voice without training it first. Every time you start a new conversation, you are starting from zero. It does not remember that your brand avoids a certain word, or that you have a specific way of describing your product’s origin story. You have to include brand voice notes in every prompt, or save a system prompt if you are using the API. On the claude.ai interface, you can use the Projects feature to maintain persistent instructions — that helps, but it requires upfront setup that most beginners skip.

It occasionally invents plausible-sounding specs. I mentioned this above and I mean it. When I gave Claude a sparse product description to work from and told it to fill out the details, it once added a specific thread count for a linen product that was never in my input. The number sounded reasonable. It was fabricated. Always cross-check outputs against your actual product data before publishing.

Long bulk runs degrade toward the end. When I asked Claude to write 10 variation descriptions in a single prompt, the first five were strong and the last three started to feel repetitive despite explicit instructions not to repeat structures. Breaking bulk tasks into groups of 4-5 produces better output than running 10 or more in a single request.

How to Build a Reusable Prompt Library Inside Claude Projects

If you are going to use these prompts regularly, do not keep copying and pasting from a notes document. Claude’s Projects feature (available on the Pro plan) lets you create a persistent instruction set that applies to every conversation inside that project.

Here is how I have set mine up:

• One project per client or per product category
• The project instructions include: brand name, tone guide, words to avoid, target audience, and platform
• Each conversation inside the project starts with just the product data — no need to re-explain the context

This setup cuts the per-listing prompt time by another 2-3 minutes because you are not re-entering the brand context each time. Over a month of active use, that adds up.

Practical Summary and Next Steps

Here is what actually matters from everything above:

• Generic prompts produce generic descriptions. The five-component structure (role, audience, product data, format, platform) is the minimum to get usable output.
• Claude handles structured, multi-part prompts better than most models, which is why it performs well for product copy specifically.
• The bulk variation prompt (Prompt #4) is the highest-leverage one for most ecommerce operators — it solves the thin content problem without requiring you to write each variant manually.
• Always include a “do not invent” instruction when your product data is incomplete. One fabricated spec can cause a real customer problem.
• Set up a Claude Project for any client or store you work on regularly. The time saved on re-entering context pays for itself within the first session.

Claude Pro runs $20/month. If you are writing more than 10 product descriptions a month, that cost is recovered in the first working day. The prompts above are free — copy them, adapt them to your product category, and build your own library from there.

If you want to go deeper on how I use Claude across my full solo business operation — not just product descriptions — subscribe to the newsletter where I publish what I actually use every week, with no vendor relationships to protect.