How AI pairing works

UpsellEngine uses Claude AI to recommend which products to pair, learning from your own store rather than a generic template.

Two signals

• Co-purchase history — which products actually appear in the same orders, and how often. This is the strongest signal.
• Catalog relationships — product types, collections, and attributes, used to reason about items that should pair even when order data is thin.

Cold start vs. established stores

• Cold start (new store or new product): before there's enough order history, UpsellEngine pairs based on catalog relationships so the placement is never empty.
• Established (roughly 50+ orders of history): UpsellEngine shifts toward co-purchase analysis, and the pairings improve as more orders come in.

Screenshot: AI-suggested pairings for a product. (TODO: add.)

Staying in control

The AI proposes; you decide. You can approve, override, hand-pick, or exclude pairings — see Manual pairings & exclusions.

Note: AI recommendations are available on the Starter plan and above; the free plan uses hand-picked offers (Plans & billing). To rank recommendations by margin, see Profit Optimizer.