An AI assistant for your online store that lets customers find products the way they would ask a good salesperson. They describe what they need, such as a sofa for a small living room with a cat, a laptop for video editing under €1,200 or a rain jacket for cycling to work. The assistant asks a follow-up question when it needs to, searches your own product feed and recommends the products that fit, with a short reason for each. Everything it shows comes from your catalog, so it never invents a product, price or spec.
Filters only work when customers speak the product language of your category: Martindale and seat depth for sofas, GPU and RAM for laptops, waterproof rating for jackets, drop and cushioning for running shoes. Most customers don't. They know their situation, not the specs that go with it. So they tick a few boxes, end up with 64 results or none at all, and leave, or email your customer service.
The assistant closes that gap. It translates the situation into the specs your catalog uses and asks when something important is missing, such as who it's for or what the budget is. Customers can keep refining in their own words: "something cheaper", "in a darker colour", "and for women?".
Furniture and home
"A sofa for a small living room, and we have a cat" → compact sofas with a scratch-resistant fabric.
Electronics
"A laptop for video editing, max €1,200" → models with enough graphics power and memory within budget.
Fashion and outdoor
"A rain jacket for cycling to work that doesn't look sporty" → waterproof, breathable jackets in a city style.
B2B and parts
"Safety shoes for warehouse staff, wide fit" → S3 safety shoes that come in a wide fit.
Live demo
The demo below uses a running-shoe shop with 100 real shoes from a retailer's product feed. Tell the assistant what you're looking for, answer its questions, and see which shoes it recommends and why.
This is a demo, not a fully built product. It shows what's possible with a limited range of 100 running shoes, and not all product information is included. Prices and discounts come from a real product feed; sizes and shop policies are simulated. In a real build, the assistant works with your full catalog and all the product information you have. Product links are affiliate links.
Your product feed
The assistant reads your catalog from the feed you already have for Google Shopping or affiliates, or straight from Shopify, WooCommerce, Lightspeed or Magento. Prices, discounts and stock update automatically.
Specs that match how people ask
Feeds often lack the specs customers choose on, such as fit, intended use or "suitable for pets". We add them once, from your product data and your own expertise.
The question becomes search criteria
The AI turns "a sofa for a small living room, and we have a cat" into a maximum width and a scratch-resistant fabric. It only chooses the criteria; your catalog decides which products match.
Results and the best picks
The customer sees the one to three best picks as product cards, with a sentence on why each fits, and can open all other matches. Then they keep talking: something cheaper, more cushioning, the women's version.
A general chatbot answers from text and can easily mention a product you don't sell, an old price or a size that's out of stock. This assistant works the other way round: the AI only interprets the question, and every product, price, discount and stock level comes straight from your catalog. That makes it reliable enough to put where customers decide what to buy: on your search bar and category pages.
Fewer dead ends
No more zero-result filter combinations. When nothing matches, the assistant relaxes the least important filter and says what it changed.
Advice like in your shop
Customers get the questions and advice a good salesperson would give, at any hour, based on your own range.
An extra way in
Customers who like filtering keep their filters. The assistant is there for everyone who doesn't know exactly what they need.
Insight into demand
Every question is logged anonymously. You see what customers look for in their own words, which products they can't find and what you might add to your range or your category pages.
The assistant is styled in your brand and can sit in your search bar, on category pages or as a "help me choose" button. It answers in the customer's language.
Intro call
We look at your store, your feed, your filters and the questions customers ask your team.
Feed and specs
I connect your feed or platform and add the specs customers choose on, for the categories that matter most.
Test with real questions
On a test page, we try the questions your customers actually ask and fine-tune the filters, the picks and the tone.
Live in your store
In your search bar, on category pages or as a "help me choose" button, with the question log in your own Google Sheet.
Monthly review
What customers ask, what they can't find and what the assistant brings in.
This assistant runs in your own webshop. Want shoppers to find your products inside ChatGPT? Start with a product feed for ChatGPT's shopping results, and consider a ChatGPT app. For a chatbot that answers questions from your content, see the AI chatbot for your website, or compare all AI integrations.
AI search on your full catalog, through your product feed or platform API
Your filters and specs mapped to how customers describe their needs
Product cards with price, discount and stock, with the best picks explained
Follow-up questions to narrow it down, like in ChatGPT
Question log: what customers look for, and what you don't sell yet
Hosting, daily feed updates and usage limits
No. The AI only interprets the question and chooses filters. Which products match, and every price, discount and stock level, come straight from your catalog. It can only recommend products from the results.
Shopify, WooCommerce, Lightspeed, Magento and any store with a product feed, such as a Google Shopping, TradeTracker, Awin or Daisycon feed in XML or CSV.
That's common. We add the missing specs once, for the categories that matter most, based on your product data, the manufacturers' specs and your own expertise. New products in those categories get the same specs.
No, it works next to them. Customers who prefer filtering keep doing that. The assistant is an extra way in, in your search bar, on category pages or behind a "help me choose" button, for customers who don't know exactly what they need.
The website chatbot answers questions from your content, such as pages, FAQs and articles. This assistant is built for finding products: it works on your structured product data and filters your catalog. Many webshops combine both, for product advice and for questions about delivery and returns.
Usually 3 to 4 weeks after the intro call: connecting the feed, adding specs for your main categories, a test week with real questions and going live. Large catalogs with many categories take longer.
Setup starts at €2,450, then from €199 per month for hosting, AI usage, daily feed updates and the monthly question report. Large catalogs, many categories, multiple languages or extra integrations are quoted separately. You get a fixed proposal after a short intro call.
Every project starts with a call or async intake to understand your goals, current workflows, and key challenges—so the solution fits your reality.
I analyze your current structure—whether that’s your content library, keyword map, editorial workflow, or AI tooling setup—and identify the highest-leverage improvements.
You get access to an online dashboard with the findings, priorities and action steps, so you can work through them one by one, plus a Loom video that walks you through the key points.
Depending on your package, I offer async feedback, follow-up Q&A, or a final strategy call to adapt the plan to your team or evolving needs.
Let's discuss how this service can help solve your specific SEO challenges.
Starting from
From €2,450 + €199/month
Get in touch to discuss your project requirements and how this service can help you achieve your goals.
Or email me directly: etienne@etiennegozems.com