AI agents don't browse your store the way customers do. They don't see your photos, your layout, or your brand story. They read text. And the first text they read is your product title.
If that title doesn't say what the product is, the AI agent moves on. It can't recommend what it can't understand.
What AI agents see
When an AI agent reads your product catalog, it gets a list of titles, descriptions, prices, and variants. No images. No context. No "you know what I mean."
A title like "The Dreamcatcher" tells the agent nothing. Is it a wall hanging? A lamp? A book?
But "Handwoven Macrame Wall Hanging, Natural Cotton, 24in" — now the agent knows exactly what it is. It can match this product to a customer asking for "macrame wall art under $50."
Title rules
Good product titles for AI follow a simple pattern: what it is + key attribute + brand (if relevant).
- Include the product type. "Shirt," "Candle," "Wall Hanging" — the most basic identifier. If the title doesn't include what the product is, AI agents can't categorize it.
- Add one or two key attributes. Material, color, size — whatever makes this product specific. "Organic Cotton T-Shirt" is better than "T-Shirt."
- Include brand name if customers search by brand. "Nike Air Max 90" is useful. For your own brand, it's less critical but still helps.
- Keep it under 150 characters. Long titles get cut off and make AI parsing harder.
- Don't go under 10 characters. A title that short can't be descriptive enough.
- Skip clever names that mean nothing out of context. "The Aurora" or "Midnight Bliss" sound nice but tell AI nothing about the product.
Before and after examples
| Before | After |
|---|---|
| The Dreamcatcher | Handwoven Macrame Wall Hanging, Natural Cotton, 24in |
| Midnight Bliss | Lavender Soy Candle, 8oz, Hand-Poured |
| The Weekender | Canvas Weekender Duffel Bag, Waxed Cotton, 40L |
| Fresh Start Bundle | Organic Skincare Starter Kit — Cleanser, Toner, Moisturizer |
Description rules
Descriptions are where AI agents look for the details they need to match products to customer queries. Specs matter more than marketing copy here.
- Include actual specifications. Materials, dimensions, weight, capacity — real numbers and units. "100% organic cotton, 24 x 36 inches, 180gsm" is what AI agents can use.
- Skip marketing fluff as the whole description. "Premium quality, handcrafted with love" tells AI nothing useful. It's fine to include some marketing language, but not as a replacement for specs.
- Add care instructions. Washing, cleaning, maintenance — these help AI agents answer specific customer questions.
- Include use cases. "Suitable for indoor and outdoor use" or "Perfect for daily commute" — these help AI agents match products to specific needs.
- Keep it over 50 characters. Very short descriptions don't give AI agents enough to work with.
- Don't rely on images alone. Some stores have descriptions that are just embedded images or iframes. AI agents can't read those — they need actual text.
Before and after examples
| Before | After |
|---|---|
| Premium quality. Handcrafted with love. The perfect gift. | 100% GOTS certified organic cotton throw blanket. 50 x 60 inches. 320gsm weight. Machine washable, tumble dry low. Available in natural, charcoal, and sage. |
| Our best seller! You'll love it. | Ceramic pour-over coffee dripper. Fits standard #2 filters. Brews 1-3 cups. Dishwasher safe. Works with mugs up to 4 inches in diameter. |
Meta descriptions matter too
When AI agents crawl your product pages (not just the JSON API), they read the meta description — the short summary that shows up in search results. If you don't write one, Shopify auto-generates it by truncating your product description. That auto-generated version is often cut off mid-sentence or pulls in irrelevant text.
A good meta description is a one-sentence summary of the product that includes the product type, a key attribute, and who it's for. Think of it as the title's backup — if an AI agent only reads the meta description, it should still know what the product is.
- Keep it between 120–155 characters. Long enough to be useful, short enough to not get truncated.
- Include the product type. Same rule as titles — say what it is.
- Add a differentiator. What makes this product specific? Material, use case, audience.
- Don't duplicate the title. The meta description should add information, not repeat it.
Before and after examples
| Before (auto-generated) | After (custom) |
|---|---|
| Premium quality. Handcrafted with love. The perfect gift. Made from the finest materials... | GOTS certified organic cotton throw blanket, 50 x 60 in. Machine washable. Available in natural, charcoal, and sage. |
| Our best seller! You'll love it. Free shipping on orders over... | Ceramic pour-over coffee dripper for #2 filters. Brews 1-3 cups, fits mugs up to 4 inches. Dishwasher safe. |
To edit meta descriptions in Shopify: open a product, scroll to the bottom, and click Edit website SEO. The meta description field is right there. It's not available in bulk edit or CSV export — you'll need to update each product individually, so start with your top sellers.
How to audit your titles right now
Here's a quick test. Open this URL in your browser:
yourstore.com/products.json?limit=10
Look at the title field for each product. Read them without any other context — no images, no collection pages, no brand knowledge.
If a title doesn't make sense on its own, it won't make sense to an AI agent either.
How to update titles and descriptions in Shopify
- Go to Products in your Shopify admin
- For individual products: click the product, edit the title and description fields, save
- For bulk updates: select multiple products, click Bulk edit, and you can edit titles in a spreadsheet-like view
- For large catalogs: export products as CSV, update the
TitleandBody (HTML)columns, and re-import
Start with your top-selling products. Those are the ones AI agents are most likely to encounter first.