How to Teach Appifire About Shipping, Returns, and Store Policies

You will teach Appifire AI Chat your shipping, returns, refund rules, and other store policies so chat answers match what you publish. You will put clear policy text into Knowledge Hub (Website knowledge and FAQ), save it into RAG, and test storefront questions, without letting chat approve refunds or invent exceptions.

When you finish, routine policy questions should get store-true answers, and judgment cases should still reach a human.

For the general automation split (policy facts vs refund decisions), see How to Automate Shipping, Returns, and Refund Policy Questions.

Direct answer

Open Appifire Knowledge Hub. Put shipping, returns, and store rules in the knowledge editor (include FAQ lines as Q: ... A: ...). Use full https:// links to your live policy pages. Click Save content so the text is ingested for chat. Test three policy prompts on the storefront. Keep refunds, damages, and “make it right” asks on your Talk-to-human path. For the full Hub setup flow, see Set Up Website Knowledge and FAQs in Appifire.

Prerequisites

• Appifire installed (Getting Started With Appifire AI Chat)
• Published shipping / returns / refund policy pages (or clear FAQ copy you can paste)
• Access to Appifire admin → Knowledge
• Storefront widget working enough to test (widget not showing checklist if the bubble is missing)
• Optional: Talk-to-human contacts set for exceptions (escalation guide)

Note: First install may auto-bootstrap website knowledge in the background. Always edit that draft, do not trust it unreviewed for legal/policy wording.

What to teach (and what not to)

Teach as facts (safe for chat)

Write clear, current rules:

• Where you ship / do not ship
• Typical processing or delivery windows (as you publish them)
• Return window and condition rules
• Who pays return shipping (as your page states)
• How to start a return (link to the form or email)
• Refund timing as policy text (“refunds are issued after…”), not “your refund is done”
• Gift cards, exchanges, international duties, if you support them

Keep human (do not automate as approval)

• Approving or denying a refund
• Exceptions outside the window
• Damaged / missing item remedies
• Chargebacks and angry complaints
• Custom “just this once” goodwill

Chat should quote the rule and hand off for the decision.

Step 1: Gather one source of truth

1. Open your live shipping policy URL.
2. Open your returns / refund policy URL.
3. Copy any FAQ answers that restate those rules.
4. Fix conflicts on the website first (product pages vs policy pages).

Expected result: One clear answer per rule before you teach the AI.

Step 2: Open Knowledge Hub

1. In Shopify admin, open Appifire AI Chat.
2. Open Knowledge Hub.
3. Confirm you see Sync Website Knowledge, the editor, and Save content.

Expected result: You can edit the knowledge base.

Step 3: Add policy text to the knowledge editor

In Review and edit your knowledge base:

1. Review any auto-fetched or installed content.
2. Delete filler or wrong lines (placeholder “not available” style text should not stay).
3. Add a clear Shipping section and a Returns / refunds section in plain language.
4. Include full https:// URLs to your live policy pages (not “Privacy Policy: Available”).
5. Keep numbers exact: days, regions, restocking fees, as published.

Optional: use Get content from website (confirm replace if the editor already has text), then edit heavily. Fetch uses AI credits / markup like other AI calls. Confirm Billing if you are on paid credits.

Expected result: The editor contains the policy facts you want chat to use.

Step 4: Add FAQ pairs for common wordings

In the same editor, under an FAQ and Common Questions heading (or similar), add lines shoppers actually ask, for example:

Q: How long does shipping take? A: …
Q: Do you ship internationally? A: …
Q: What is your return window? A: …
Q: Who pays for return shipping? A: …
Q: How do I start a return? A: …

Keep answers short and aligned with your live policy pages.

Expected result: Paraphrased questions still map to the same rules.

Step 5: Save so chat can use the text

1. Click Save content.
2. Wait for Content saved successfully (retry if you see an embeddings warning).
3. Do not test on an unsaved draft.

Saving stores the text and ingests it into knowledge chunks for retrieval.

Expected result: Knowledge is persisted and available to chat.

Step 6: Test on the storefront

Ask these on the live widget (private window):

1. “How long does shipping take to [your top region]?”
2. “What is your return policy?”
3. “Do you offer refunds after [N] days?”
4. “Please refund my order now.”
5. “Talk to a human.”

Pass rules

• Answers 1-3 match your published pages.
• Answer 4 does not invent a completed refund; it explains policy or hands off.
• Answer 5 gives a real contact path.

Expected result: Policy facts work; judgment fails safe.

Step 7: Keep policies fresh

When you change shipping rates or return windows:

1. Update the live Shopify pages first.
2. Update Knowledge Hub Website knowledge and FAQ the same day.
3. Save again.
4. Re-run the three policy tests.

Stale Knowledge Hub text creates confident wrong answers even when the website is fixed.

Verification checklist

• Live policy URLs open and match what you intend to promise
• Knowledge editor includes shipping + returns sections with full https:// links
• FAQ Q&A lines cover the top paraphrases
• Save content completed / ingested
• Storefront tests match the pages
• Refund-request test does not fake a refund
• Talk-to-human path works
• Owner named for future policy edits

Common failures and fixes

Symptom	Likely cause	Fix
Chat invents a 30-day window	Knowledge empty or generic bootstrap left unedited	Paste your real rules; save
Chat conflicts with the policy page	Hub text outdated	Sync Hub to the live page; save
“Refund issued” style answer	Over-broad FAQ / bad prompt expectation	Rewrite FAQ; retest; escalate path
No policy answers at all	Not saved / no chunks	Save content again; confirm editor is not empty
Good policy, bad product facts	Different lane	Fix catalog sync, not policies (product info guide)

Limits and non-goals

• Knowledge Hub teaches store knowledge; it does not process returns in Shopify Admin.
• Appifire does not replace a helpdesk for disputes.
• Auto-bootstrap and fetch can miss or blur legal detail, you own the final wording.
• Product-specific shipping exceptions should also appear on the PDP when they matter.
• Order status is a separate live lookup path (order status in Appifire).

Related Appifire guides

• How to Automate Shipping, Returns, and Refund Policy Questions
• How an AI Shopping Assistant Uses Product, Policy, and Store Knowledge
• Improve Product and Order Answers in Appifire AI Chat
• Shopify Customer Support Automation: What to Automate and Keep Human
• Getting Started With Appifire AI Chat

Last verified: 2026-08-09 against Appifire Knowledge Hub behavior (Website knowledge + FAQ, first-install bootstrap, fetch/edit/save ingest, sanitizer/URL guidance) in product docs. Re-check admin labels after releases.

Next action

Open Knowledge Hub today. Paste your shipping and returns rules with live https:// links. Save. Run the five storefront tests above. Fix any mismatch on the website first, then in Knowledge Hub, never only in chat macros.