Shopgentic · Playbooks
Automate your product uploads with AI: the complete playbook
What you'll have when done
A workflow where a supplier sheet goes in and a drafted Shopify product page comes out: title, description, SEO meta and image alt text, all in your tone of voice. You review and publish. Drafting time per product drops from 20 to 30 minutes to just a few.
Written by Daan · Commercial & Creative
Builds and runs these systems at Shopgentic. Updated July 2026.
The manual version, costed honestly
A typical product upload: read the supplier sheet, write a title and description, translate specs into benefits, write the SEO meta, name and alt-text the images, fill in variants. Done properly that is 20 to 30 minutes per product. At 15 new products a week, which is normal for a growing store, that is 6 to 8 hours of skilled work, every week, forever. The person doing it is usually the same person who should be doing marketing.
Nothing held back, this is the whole loop
In
Supplier sheet, spec table or a few photos land in a shared folder
Draft
Claude drafts copy, meta and alt text using your tone document as context
Stage
The result becomes a draft product in Shopify, never published directly
Approve
A human reads it, fixes what’s off, hits publish. That judgment stays yours
First teach it your voice, then hand it the work
Create a Claude Project called “Product pages”. Add your five best existing product descriptions, your size charts and your returns policy as project knowledge. Then run these two prompts.
Prompt 1 · extract your tone, run once
Here are 5 product descriptions we wrote ourselves and are proud of. Analyze them and write a tone-of-voice guide for our store: sentence length, warmth, how we talk about materials and fit, words we use, words we never use. Keep it under 300 words. I will use this guide to brief you on future products.
Prompt 2 · the draft template, run per product
Using our tone-of-voice guide, draft a Shopify product page from this supplier data: [paste spec table or supplier text] Deliver exactly: 1. Product title (max 60 characters) 2. Description (120–180 words, benefits before specs) 3. SEO title and meta description 4. Alt text for each image: [list the images] 5. Three questions a customer would ask, answered in one line each If a spec is missing or ambiguous, ask me instead of guessing.
That last line is the most important one in the whole playbook. An agent that guesses specs will eventually publish a wrong washing instruction. An agent that asks stays trustworthy.
The part nobody mentions
This setup decays if nobody owns it. Once a month, swap the weakest example description in your project knowledge for the best one you published recently, so the tone keeps improving instead of drifting. Give the workflow one owner, not a rotation. And when you catch yourself heavily editing every draft, fix the tone guide, not each draft. Ten minutes of maintenance a month keeps the three-minutes-per-product number real.
Results to expect
Expect the drafting itself to drop from 20 to 30 minutes per product to a few, once the tone guide is dialed in. The review step stays, that is the point. Count on a week or two of correcting the tone guide before the drafts read like yours, and budget real time for that, it is where the quality comes from.
Where this breaks
Never auto-publish
Everything lands as a Shopify draft. Publishing is a human click, always.
Price and stock are human-only
The agent drafts copy. It does not touch pricing, inventory or variants.
Flag, don't guess
Missing or contradictory specs come back as questions, not as invented facts.
Related playbooks
How do Shopify stores automate customer support with AI agents?
Shopify stores automate support by connecting an AI agent to their inbox, helpdesk and store data. The agent…
Set up an AI agent that answers your support email — in 7 steps
Before: every support email is written from scratch, 5 to 10 minutes each. After: a drafted reply in your…
Monthly, no lock-in. You own the system.