Read the press release | Marqo Sibbi overview
Marqo dropped Sibbi in May 2026. It's a conversational commerce agent that handles the entire shopper journey in one interface. Natural language product descriptions. Photo uploads for visual search. Personalized recommendations. Cart transactions. Returns initiation. Same conversation. No passes to chat support. No "visit this page." No friction.
This is the architecture your ecom business should obsess over.
The Problem That Won't Die
Your search bar is broken. You know it. Your shoppers know it.
94% of U.S. consumers have landed on irrelevant search results. 74% have seen text-mismatched returns. 81% of U.S. shoppers leave the site after one bad search experience—globally that's 80%. More than half your sites deliver mediocre or worse search UX. Traditional ecommerce search converts at 12%, and 40% of queries return irrelevant results.
It's not a feature problem. It's an architecture problem. Search is a dead format when customers know what feeling they want but not what SKU to type.
What Sibbi Does Differently
Sibbi removes the search bar entirely. Replaces it with conversation.
Instead of "men's running shoes under $150," a shopper says "I need something that feels like lightweight New Balance but with Adidas cushioning." Or uploads a product photo and says "something like this but in navy." Sibbi understands visual attributes, semantic relationships, catalog inventory, and commercial signals, margin, availability, personalization data tied to that specific customer.
It delivers guided discovery (not keyword matching), cross-sell recommendations, and direct cart addition, all in the same flow. Post-purchase? Shoppers ask "where is my order" or initiate a return in the same conversation. No channel switch. No support ticket. No friction.
The math behind it: Marqo built Commerce Superintelligence, a single intelligence layer trained on each retailer's specific catalog, grounded in real-time inventory, and trained to recognize visual and semantic patterns. Every retailer gets their own AI model. That's the difference between commodity search and owner-operated search.
The Receipts
KICKS CREW saw a 17.7% conversion rate improvement. Mejuri generated a 19.8% jump in search-driven conversion. Fashion Nova attributed $130 million in incremental revenue to Marqo's implementation. SwimOutlet went live in five days and hit a 10.6% increase in search add-to-cart rate. These numbers are validated through production A/B tests, not projections.
Conversational commerce is projected to reach between $43 billion (by 2028, per Juniper) and $290 billion (longer projection, same source). The market is moving. The question is whether you move with it or defend the broken search bar.
How It Works: Integration and Setup
Sibbi deploys with a single line of code. The platform surface measurable results within 14 days. Marqo offers pre-built connectors for Shopify, Adobe Commerce, and Salesforce Commerce Cloud. Sibbi and Commerce Superintelligence are available for enterprise retailers via API. Pricing is custom (contact Marqo directly for your business profile).
The technical requirements are minimal, if you're already running a modern ecom stack, you can plug this in. The real requirement is organizational skin in the game: you need to measure the impact, iterate on recommendations, and give the agent permission to own the discovery moment.
The Amazon Elephant
Amazon's Alexa for Shopping does something similar, voice-driven discovery, cart integration, post-purchase support. Same vision. Different scale. Amazon's advantage: massive user base and unified identity across devices. Sibbi's advantage: your brand controls the model, the recommendations, the economics. Amazon takes its cut. With Sibbi, the upside stays with you.
That's the trade-off: less distribution, more control.
Search Is Where Intent Lives
In 1997, AIN built its own search for a $100M+ catalog, before Google dominated, before APIs. The lesson was simple: whoever owns the moment when a customer articulates intent owns the conversion. Search wasn't a feature to outsource. It was the core property.
Twenty-nine years later, the principle hasn't changed. The delivery method has. Instead of keyword boxes, it's conversation. Instead of broad matching, it's semantic understanding. Instead of generic algorithms, it's your data, your catalog, your model.
Systems beat slogans. A conversational system that delivers 17%+ conversion gains beats a search bar that delivers 40% irrelevant results. That's not opinion. That's math.
One Caveat
Conversational commerce adoption depends on shopper behavior change. Not every shopper wants to chat. Not every category works equally. Fashion and home goods? Perfect fit. Commoditized or lowest-price categories? Harder sell. Sibbi will work best for premium and differentiated brands where recommendations matter, where the conversation is actually valuable, not another friction point.
The agent also needs training. You can't drop it live, ignore it, and expect it to perform. Like any system with skin in the game, it requires operational rigor: measuring what works, pruning what doesn't, iterating on recommendations.
What You Do Next
Don't wait for perfect. Get on Marqo's roadmap. Run a pilot with Sibbi, measure conversion, cart value, shopper sentiment. Build in measurement from day one. If Fashion Nova, KICKS CREW, and SwimOutlet saw measurable upside, your business probably will too.
If your search bar is still the front door to discovery, you're behind. The next era of ecommerce isn't about better keyword matching. It's about owning the moment when intent enters the system, and owning it in a format that actual humans prefer.
Conversation beats search. The data says so. The receipts say so.
FAQ
Q: Does Sibbi replace my entire search infrastructure? A: For most retailers, yes. It handles product discovery, recommendations, cart, and post-purchase. Some brands keep a secondary traditional search for power users or bulk orders. For most ecom, Sibbi becomes the primary discovery engine.
Q: How does Sibbi handle inventory accuracy? A: It's grounded in real inventory. If you're out of stock, Sibbi knows it and can recommend alternatives or notify when items are back. Real-time sync is standard.
Q: Will my customers actually want to chat instead of search? A: Early data says yes, when the conversation is fast and personalized. Fashion, beauty, home goods show adoption rates of 25-40% of shoppers using the conversational layer within the first 30 days. Lower in commodity categories.
Q: What if I'm on WooCommerce or a non-standard platform? A: Marqo's official connectors cover Shopify, Adobe Commerce, and Salesforce. Custom integrations are possible via API, but you'll need engineering resources or a partner to build the connector. Worth asking Marqo about your specific stack.
Q: What's the upside for a $3-5M revenue store? A: Comparable retailers saw 15-20% conversion gains. For a $5M store at 2% AOV conversion, a 17% improvement could mean $127K additional annual revenue. The payoff is real even at smaller scale, if you execute the setup and measurement right.
*Disclosure: Jeff Barnes, MBA has no personal position in any company, fund, or platform named in this article. demg.ai has no current commercial relationship with any party mentioned. demg.ai provides marketing education and consulting services, not investment advice. Past performance does not guarantee future results.*