AI product finder for Shopify and WooCommerce that lifts ATC 34%.
Keyword search returns 412 results. Zubby turns vague intent — 'a gift for my dad who likes hiking' — into the three best products from your live catalog, with reasoning, in under a second. Add to cart from inside the chat. 34% add-to-cart lift, 1.8x pages per session.
- +34%
- add-to-cart lift
- 250k+
- SKUs supported
- <300ms
- p50 retrieval
- 25
- languages auto-detected
Built for Shopify and WooCommerce·See integrations
What is AI product discovery?
AI product discovery is a conversational interface that helps shoppers find the right product by intent, context, and constraints — not by guessing the right keyword. Instead of a search bar that returns 412 results, the shopper has a 4-turn conversation that ends with the top three matches and a one-tap add-to-cart. The agent asks the clarifying questions a knowledgeable salesperson would, and only recommends products you actually carry.
Under the hood, every product, variant, and policy is embedded into a 1536-dimensional vector space and joined at query time with live inventory and price data. When the shopper says 'something warm for the office that isn't too corporate,' the agent retrieves on intent — not literal token match — then layers structured filters ('just the wool ones under $300') on top of the semantic result. p50 retrieval is under 300ms on catalogs up to 250,000 SKUs.
The economic difference vs static search is large: Zubby stores see a 34% lift in add-to-cart, 1.8x more pages per session, and 63% self-serve resolution on product questions. Vision input on paid plans lets shoppers upload a photo and the agent matches your catalog by visual + semantic similarity — useful for furniture, fashion, beauty, and anything where look matters more than category.
Three turns, one decision.
From a vague question to a confident add-to-cart, with the same logic a great salesperson uses. See the knowledge base feature or read the Shopify install guide.
Shopper asks
In their own words. 'Something for a weekend in Maine' or 'a moisturiser for combo skin under $40'.
Agent clarifies
One or two targeted follow-ups — never a 12-question quiz. The agent only asks what genuinely narrows the result.
Top three, with reasoning
Three best matches as visual product cards with one-tap add-to-cart. Each card explains why it's a fit, citing live stock and price.
Built for the way shoppers actually decide.
Every product, policy, and variant indexed. Every conversation tuned to your store's voice.
Vector + structured retrieval
1536-dim embeddings on every product, blended with rules-based filters (price, attribute, stock) for grounded recommendations.
Mid-conversation filters
'Just the waterproof ones' applies a structured filter on top of the semantic result — context survives across turns.
Clarifying-question engine
The agent asks only when ambiguity matters. No quiz fatigue, no irrelevant follow-ups.
In-chat product cards
Visual cards with photo, price, variants, and one-tap add-to-cart — no detour to the listing page.
Reranking that learns
Weekly retraining on real merchandising signals: clicks, adds, purchases, returns. Your catalog gets smarter over time.
Vision-aware discovery
Shoppers can upload a photo of a vibe or a competitor product; the agent returns the closest matches you carry.
From 'I have no idea' to add-to-cart in 30 seconds.
Real outdoor brand, real cold-weather query. Search-only stores lose the shopper at the filter step.
Without Zubby
- Shopper types 'warm jacket' and gets 187 results.
- Filters take three clicks; some aren't in stock.
- Returns to search, adjusts query, retries.
- Tab closes 90 seconds later. No purchase.
With Zubby
- ShopperI need something warm for a weekend in Vermont.
- ZubbyGot it. Will you be outdoors most of the time, or balancing outdoor + indoor? And what's your size?
- ShopperMostly hiking. Size M.
- ZubbyHere are three I'd recommend — all in M, in stock, rated for sub-freezing. The Alta Parka is the warmest, the Ridge Shell is the most packable.
Real outcomes on real stores.
Across Zubby stores that replaced static search with conversational discovery. Full method note on the Benchmark Report.
+34%
Add-to-cart lift
Median lift after replacing static search with Zubby discovery.
1.8×
Pages per session
Conversational discovery keeps shoppers engaged longer.
63%
Self-serve resolution
Of product questions answered without a human.
<300ms
p50 retrieval
Sub-second from question to top-three recommendations across 250k SKUs.
Source: Zubby Q2 2026 benchmark data, n=14,000+ active stores.
Conversational discovery vs classic search.
- Input styleKeyword guessNatural language intent
- Result countHundreds, ranked by relevance scoreTop 3 with reasoning
- Stock awarenessOften shows out-of-stockLive inventory grounding, always
- ClarificationFilters and facetsTargeted follow-up questions
- Add-to-cart frictionClick → PDP → addAdd from inside the chat
- Learning loopStatic indexWeekly relevance retraining on real signals
Built for Shopify and WooCommerce — more platforms coming
- SOC2 Type II ready
- GDPR + CCPA workflows
- 99.97% uptime SLA
- EU + US data residency
Product discovery, answered.
Everything the procurement, security, and ops teams ask before they roll us out.
AI search returns a list of 412 results ranked by relevance score. AI product discovery has a conversation: it asks one or two clarifying questions (occasion, budget, fit), narrows the catalog, surfaces the top three with reasoning, and lets the shopper add to cart from inside the chat. Search is one tool the agent uses — not the whole experience. Zubby stores see a 34% lift in add-to-cart and 1.8x more pages per session.
Related use cases
Revenue rescue
AI cart abandonment recovery
Detect hesitation in real time, intervene before the tab closes, and follow up with on-brand recovery email and SMS.
ExploreAround the clock
24/7 AI customer support
Tier-1 questions answered in under a second — order tracking, returns, sizing, policies, and stock.
ExploreLift AOV
AI upsell + bundle recommendations
Complementary products surfaced in conversation, plus dynamic bundles that respect inventory and margin rules.
ExploreGive your shoppers a finder, not a search bar.
14-day trial on Shopify and WooCommerce. Install in 10 minutes. First guided session usually in the first hour of traffic. See pricing or compare to your current search.