Teaching Your Sifu: How It Learns Your Store

The difference between a chatbot that frustrates shoppers and one that closes sales comes down to a single question: does it actually know your store?

ShopSifu is built to know it. Instead of inventing answers, Sifu studies your real catalog and your own documents, then responds from them — and tells the shopper where the answer came from.

Two sources of truth

Sifu learns from two places:

1. Your live Shopify catalog — products, variants, prices, and availability, synced through the Shopify Admin API. When a shopper asks "is the medium in stock?", Sifu is reading your real inventory, not a stale snapshot.
2. Your documents — the policies and guides that live in your team's head and your help pages: shipping, returns, sizing, warranty, care instructions, FAQs.

Why grounded answers matter

A general AI model is fluent but uninformed about your shop. Ask it your return window and it will happily guess — confidently and wrong. That is worse than no answer.

Sifu answers from your retrieved content, so:

• It cites the policy it used, instead of improvising.
• When it doesn't know, it says so and hands off — rather than hallucinating.
• Update a doc, and Sifu's answers update with it. No retraining, no waiting.

What to give it on day one

Start with the documents that drive the most questions:

• Shipping times and costs
• Returns and exchanges
• Sizing or fit guides
• Product care and warranty
• Your top 10 FAQs

A good rule: if a shopper has ever emailed to ask it, Sifu should be able to answer it. Feed it those emails' answers.

It gets sharper over time

As you add documents and refine your tone, Sifu's answers get more precise and more on-brand. Training your sifu isn't a one-time setup — it's a short, occasional habit that compounds.

See what Sifu can learn from your store. Start free.