AI Product Search Guide

How Velaro's AI finds products in your store

When connected to Shopify, BigCommerce, WooCommerce, or Magento, Velaro automatically reads your entire product catalog and builds a searchable knowledge layer. The bot can find products by meaning — not just exact keyword matches. A visitor asking "do you have anything good for sore knees?" will find knee braces and compression sleeves even if those exact words aren't in the product listing. This works because Velaro understands product intent, not just text.

The catalog is indexed when you first connect your store and refreshed automatically on a daily schedule, so new products appear without any action on your part.

Will my bot show the wrong price if prices change?

No. Prices are always fetched live from your store at the moment a customer asks. They are never stored in the knowledge index. If you change a price at 9am, a customer asking at 9:01am sees the new price. The same is true for inventory — the bot always asks your store "is this in stock right now?" in real time.

This is a deliberate design choice: product descriptions and names are stable and worth caching for fast search, but prices and stock levels change constantly and must always be live.

What makes product search accurate?

Two search methods run simultaneously on every query and their results are merged:

1. Semantic search — finds products by meaning, handles synonyms and natural language.

2. Keyword search — exact matches on product names, SKUs, and model numbers.

The merged result gives both precision (exact SKU lookup) and recall (intent-based discovery). Neither method alone is as accurate as both together.

How to improve your bot's product search accuracy

Better product data means better search. Tips:

1. Write real descriptions — even 2–3 sentences dramatically improve how the bot finds products. Products with no description are hard to surface via intent queries.

2. Use descriptive variant names — "Small / Red" is better than "Option A".

3. Add relevant tags in Shopify or BigCommerce — tags are indexed and used for discovery.

4. Set clear product types and categories — helps the bot narrow results when a customer describes a category rather than a specific product.

5. Keep product titles clear — "Men's Trail Running Shoe — Waterproof" outperforms "TRS-X400-M".

What content does the bot search when answering questions?

The bot searches all your enabled knowledge sources simultaneously in one query:

• Your product catalog (from your connected store)
• Your published KB articles
• Your website pages (if you have run a site scan)
• Any uploaded documents (PDF, Word)
• Resolved conversations that your agents have answered well

All sources are searched at once — the most relevant answer wins regardless of where it lives. You do not need to tell the bot which source to check.

Which stores are supported for product search?

Full automatic product indexing and AI search is available for: Shopify, BigCommerce, WooCommerce, Magento/Adobe Commerce, and Square. Products are indexed on a daily schedule.

To trigger an immediate re-index after a major catalog update, go to Integrations → [your store] → Bot Knowledge → Sync Now.

What products won't the bot find?

Draft products, archived products, and hidden products are not indexed — only published, active products. Products with no title cannot be meaningfully indexed. Products set to a sales channel that excludes your storefront may also be skipped depending on your store's API permissions.

If a product isn't showing up in bot answers, check that it is published and active in your store admin.