AI Product Copy and the Smart Offer Engine

Overview

StretchShop includes two AI-assisted capabilities that help you sell more and work faster: AI product copy, which drafts marketing text for a product, and the smart offer engine, which powers product recommendations, dynamic offers, and checkout risk scoring. Both are optional and may need to be enabled for your workspace; if you don't see them, contact your administrator.

AI product copy

Instead of writing every listing from scratch, let StretchShop draft it.

1. Open a product (or the new-product form) and find the Generate copy action.
2. Provide the product name and any attributes you know — material, scent, size, audience.
3. Run it. The assistant returns a suggested title, a compelling description, feature bullets, and a set of SEO keywords.
4. Review and edit the draft, then apply the parts you like to the product's fields.

The AI is a drafting aid: it's fail-safe (if it's unavailable, your product simply keeps its existing copy) and always leaves you in control of the final wording.

The smart offer engine

The engine analyzes a cart and catalog to help you convert and protect each sale. It offers three functions:

• Recommendations — given the current cart items and your active catalog, it suggests complementary products to cross-sell ("customers also buy").
• Offers — given the cart's subtotal, item count, and currency, it suggests a dynamic offer (such as a threshold incentive) to nudge the buyer toward completing or growing the order.
• Checkout risk — given signals like the customer, whether billing and shipping match, and recent order velocity, it returns a risk assessment so you can add friction or review suspicious checkouts.

These run as a stateless compute service: StretchShop sends the current cart and catalog inputs and uses the result — nothing about the engine changes your stored data on its own.

A realistic example

You add a new candle and click Generate copy; the assistant proposes the title "Cedar & Sage Soy Candle — Hand-Poured, 45-Hour Burn," a warm description, four benefit bullets, and keywords like "soy candle," "cedar sage," and "hand poured." You tweak a line and save. Later, a shopper has one candle in their cart; the recommendation engine suggests a matching wick trimmer, and the offer engine proposes "Add $16 more for free shipping." A high-value order with mismatched billing/shipping and several orders in the last hour comes back with an elevated risk score, so you place its fulfillment on hold for review.

Tips

• Treat AI copy as a first draft — edit for brand voice and accuracy before publishing.
• Give the assistant specific attributes; richer input yields sharper titles, bullets, and keywords.
• Use recommendations on cart and product pages to lift average order value.
• Feed offer suggestions into free-shipping thresholds and bundle prompts.
• Use checkout risk to gate only the riskiest orders — pair a high score with a fulfillment hold rather than blocking everyone.

Troubleshooting

• "Generate copy" isn't available — AI copy may not be enabled for your workspace; ask your administrator.
• Copy came back empty — the model may be temporarily unavailable; the product keeps its current copy, so simply try again later.
• Engine features are disabled — recommendations, offers, and risk require the offer engine to be enabled; if calls report the engine is off, contact your administrator.
• Recommendations look generic — they draw on your active catalog and the current cart; a richer, well-tagged catalog produces better suggestions.