AI assistant for e-commerce
I help customers choose from your catalog, compare variants and pin down what they need. Availability comes from your connected data; order status and return requests go through your configured APIs.
Typical pain
Shoppers get lost in hundreds of SKUs, generic bots invent products, and support keeps answering shipment questions. Sales chat should sell from a real catalog — not make things up.
- Recommendations outside the assortment or with wrong prices
- No variant comparison (“why this one?”)
- Order status only via email or phone
- Returns without structured intake create chaos
What I handle
- Search and recommend only from a real Heureka / product feed
- Compare products and explain why this one
- Check stock, shipping, and variant availability
- Order status via shop API (API)
- Returns or claims → helpdesk ticket
- Size, care, and invoicing FAQ from knowledge
Outcomes to expect
Product cards only from your feed — no invented SKUs or prices.
Shopping advice grounded in the catalog plus a clear next step.
After purchase: self-serve shipment status or a structured claim without repetition.
A typical shopping chat
I understand the need
I clarify budget, use case, size, or compatibility — no random tips.
I recommend from the catalog
Cards come only from your feed: real price, stock, product page link. I don't invent availability.
I add the policy
Shipping, returns, or size charts from the knowledge base — with citations.
I help after purchase
Via a connection I check the order, start a return, or hand off to a specialist.
Concrete scenarios
Choosing between two products
The shopper hesitates between two SKUs. I compare feed attributes, explain differences, and link to the product page — without invented claims.
“Do you have size M in stock?”
I check variant availability in the catalog or via API. If not, I suggest an alternative from the same feed or a restock note from knowledge.
Claim after delivery
Cited returns policy + form (reason, order ID) + helpdesk ticket. Edge cases go to a human.
“What does shipping to Brno cost?”
I answer from shipping policy in the knowledge base and add an estimate by cart / region. If data is missing, I offer a form or hand off to an operator.
A customer calls with a product question
The customer describes what they need over the phone. Based on the parameters, I recommend a specific product from the catalog — not just point them to the website.
A customer calls before driving to the store
Before heading to a physical store, the customer calls to check if an item is in stock at a specific branch. I verify availability via API and give the opening hours too.
What it looks like in practice
Concrete examples of conversations with the assistant in this industry.
A customer looks for a product by need. The assistant recommends only items from the real feed — with price, availability, and a link to the product page.
Browsing the catalogue
A customer looks for a product by need. The assistant recommends only items from the real feed — with price, availability, and a link to the product page.
A question about a specific size and stock. The assistant checks variant availability and suggests an alternative when needed.
Size in stock
A question about a specific size and stock. The assistant checks variant availability and suggests an alternative when needed.
Filtering dresses by colour and size from the catalogue — no invented SKUs.
Colour and size
Filtering dresses by colour and size from the catalogue — no invented SKUs.
When the preferred variant is unavailable, the assistant suggests similar storage boxes from the same feed.
Alternative products
When the preferred variant is unavailable, the assistant suggests similar storage boxes from the same feed.
Asks about the return and guides the customer through the store’s return process.
Return request by phone
Asks about the return and guides the customer through the store’s return process.
Browsing the catalogue
A customer looks for a product by need. The assistant recommends only items from the real feed — with price, availability, and a link to the product page.
Real deployments
See how Kanbu helps our customers in practice.
Example connections
These are examples of connections supported by your systems. Configure API or MCP tools and assign them to the assistant; knowledge sources and product feeds provide content for answers.
Catalog, prices, stock, in-chat product cards
Order status, shipping, shipment tracking
Create return or claim ticket with context
Return and shipping policies, size charts
Exchange address, preferred delivery window, confirm details
Pass to sales or support specialist with the full conversation history
Useful knowledge sources
Keep the product feed separate from the knowledge base. Knowledge is for policies and guidance — not the whole catalog as text.
- Shipping, returns, exchanges, invoicing
- Size charts and product care
- Brand story and buying guides (playbooks)
- Website or Drive sync for policy updates
Today and coming soon
Today: product feed, API connections, website and phone. Coming soon: native plugins for shop platforms and more channels (e-mail, WhatsApp, Instagram).
Try the e-commerce assistant
Try answers from your website and add a product feed for recommendations. Configure a connection to your system for order-related actions.
