Most AI-generated product descriptions read like a thesaurus exploded — generic, keyword-stuffed, and invisible to Google. This guide gives e-commerce professionals a repeatable system to prompt ChatGPT, Claude, or Gemini for descriptions that rank in search and convert browsers into buyers.

Why most AI product descriptions fail at SEO

Vague prompts → vague output that no specific shopper is searching for.
No structured data → Google can't surface the page as a rich result.
Stuffed keywords → readability tanks, bounce rate spikes, rankings drop.
No buyer intent → the copy describes features instead of solving a job-to-be-done.

The CARE framework

CARE is a four-part prompt scaffold that consistently produces SEO-optimized product copy:

C — Context. Who is the shopper, what problem are they solving, what is the brand voice?
A — Attributes. The literal facts: materials, size, weight, ingredients, warranty.
R — Ranking signals. Primary keyword, 2–3 secondary keywords, semantic variants, and the search intent (informational, commercial, transactional).
E — Experience. The output shape: H1, 2-sentence hook, 5 benefit bullets, 3 FAQs, JSON-LD Product schema.

The CARE prompt template

You are a senior e-commerce copywriter writing for [BRAND], a [CATEGORY] store targeting [BUYER PERSONA] in [MARKET].

Product attributes: [paste spec sheet — name, materials, dimensions, price, warranty, use cases].

SEO ranking signals: Primary keyword: "[PRIMARY KEYWORD]". Secondary: "[KW2]", "[KW3]". Search intent: [transactional / commercial]. Avoid keyword stuffing — use each phrase only where it reads naturally.

Output:
1. H1 (≤60 chars, contains primary keyword).
2. Meta description (≤155 chars, contains primary keyword + one secondary).
3. Two-sentence emotional hook (buyer's problem → outcome).
4. Five benefit-led bullets (benefit first, feature second).
5. Three FAQs answering the most common pre-purchase objections.
6. JSON-LD Product schema (name, description, brand, sku, offers).
7. 3 internal link anchors pointing to related collection pages.

Worked example: organic cotton t-shirt

Drop the CARE template into ChatGPT, Claude, or Gemini with the spec sheet and the primary keyword "organic cotton t-shirt for sensitive skin". In one pass you get a draft that:

Uses the primary keyword in H1, meta, and the first 100 words — once each.
Leads each bullet with the buyer benefit ("Stops the night-time itch that wakes you up") and follows with the feature ("GOTS-certified organic cotton, no dyes").
Ships a Product JSON-LD block ready to paste into the product template — eligible for Google's rich result.
Includes 3 FAQs that Google can surface as a FAQPage snippet, doubling your SERP real estate.

QA checklist before publishing

Primary keyword appears in H1, meta description, first 100 words, and one image alt — and nowhere else forced.
Readability ≥ grade-7 (paste into Hemingway).
Product + FAQPage JSON-LD validates in Google's Rich Results Test.
Page has 3+ internal links to related collections or guides.
At least one user-style review or quote (E-E-A-T signal).

Scaling to 1,000 SKUs

Save the CARE template once, then feed it your product catalog as a CSV in batches of 20. Most teams move from "8 minutes per description" to "8 seconds per description" — at higher SEO quality than the manual version.

Get the full prompt pack

The CARE template is one of 50+ copy-paste prompts inside The AI Productivity Playbook — including category page templates, FAQ generators, internal-link prompts, and the JSON-LD schemas for Product, Article, FAQPage, and BreadcrumbList. Upgrade to Pro for the full library plus the 30-day adoption plan.

How to Prompt AI for SEO-Optimized Product Descriptions (CARE Framework)