Skip to content

November offer: 50% off your November fee on monthly plans, or off a full year on yearly plans.

View Plans

How AI Product Recommendations Work in Chat

Customers describe what they want in their own words; AI turns that sentence into the right product from your catalog, with an image and an order link. We walk through each step of that chain, how an AI that works like a sales assistant differs from a rule-based bot, how to prepare your catalog and what drives the cost.

A customer types "black casual sneakers, size 42" into the chat. On your website, finding that takes three filters and some scrolling; in chat, one sentence is enough. This is where an AI product recommendation system comes in: it understands the request, checks the catalog and puts a matching product, with its image, in front of the customer. Below we explain how this chain works, how to prepare your catalog and what determines the cost.

What does an AI product recommendation system do?

A website filter expects customers to speak your language. They pick a category, tick a color box, find their size. AI does the opposite: it takes the customer's own sentence and translates it into products from your catalog.

Think of the AI doing this job as a sales assistant in a store. A bot that answers FAQs replies to "how many days do I have to return an item?" A sales assistant asks "where will you wear them, what's your budget?", narrows down the options and makes the decision easier. The two jobs may look alike, but they serve different purposes.

How does "black, casual, size 42" become a product?

In chat, AI-powered commerce (AI commerce) comes down to a five-step chain. It helps to look at each step from both the customer's side and yours.

1. Understanding the request

The AI pulls out the color (black), use (casual), size (42) and category (sneakers) from the sentence. If the customer types something misspelled or abbreviated like "sz 42 sneekers", that's fine; it tolerates typos and abbreviations. If information is missing, it asks: for running or for everyday wear?

2. Catalog search and stock check

Next it goes to the catalog. Stock and price data are updated at regular intervals from your connected store, and recommendations are based on that data. So instead of pushing a product that doesn't show as in stock, it recommends what you actually have.

3. Recommendations with images, plus alternatives

Customers don't decide without seeing the product. Each recommendation comes with a product image and a short description; if needed, a PDF catalog can be sent too. When there's no exact match, it offers close alternatives rather than sending the customer away empty-handed.

This is where context tracking matters. When the customer asks "do you have this in white?", the AI remembers which product they mean. No need to start over.

4. Product link and order

Once the customer decides, the AI directs them from the catalog to an order and payment link. Payment is completed through that link. The chat takes the sale to the door; walking through it is up to the customer.

5. Handing over to the sales team when needed

Not every sale should be left to a bot. A business customer asking for a bulk order, someone asking for special pricing, a buyer who can't decide and keeps asking questions... In these cases the AI hands the conversation to an agent along with the chat history. The agent doesn't ask "what were you looking at?" from scratch, but picks up where things left off. Sentiment analysis also instantly tags angry or worried customers, so your team doesn't miss them.

Stock is updated at regular intervals, not in real time For products that sell out quickly during the day, there may be a brief gap between your store's stock and what the chat shows. For items down to their last few units, make it a habit to confirm stock before approving the order.

Rule-based chatbot or AI sales assistant?

These two are often confused. In a rule-based flow, the customer chooses from buttons and you map out every step in advance. With AI product recommendations, the customer writes freely and the AI reaches the product by asking about their needs.

CriterionRule-based flowAI product recommendations
How does the customer write?Chooses from buttons and listsFreely, in their own words
What it does bestOrder forms, appointments, surveys, return scenariosAsking about needs, narrowing options, offering alternatives
SetupSteps are mapped out in the flow designerCatalog and knowledge base are fed in
Who is it for?Businesses with few products and a clear processStores with many products and lots of color and size variants

In Chatopya, these are two separate capabilities. With the drag-and-drop flow designer, you build flows such as order forms or return scenarios without code. On the product recommendation side, the AI takes on the role of a sales assistant: it asks about needs, recommends suitable products from the catalog and speaks multiple languages, Turkish first among them. Details are on the WhatsApp customer service and AI chatbot page.

In practice, we see the best results when both are used together. The AI finds the product, the flow collects the return request.

Try it with your own catalog. Let's look together in a demo at how product recommendations work with your products.

Book a free demo

How do you prepare your catalog for accurate recommendations?

The AI knows what your catalog says. Don't expect it to guess that a product described as "Model X, black" is suitable for everyday wear. In short: recommendation quality starts where data quality starts.

  • Descriptions: State color, material and intended use (casual, running, office) clearly.
  • Sizes: Enter them the same way for every product. Writing "42" on one product and "EU 42" on another makes matching harder.
  • Images: Choose clear photos that show the product on its own.
  • Stock and price: Keep them up to date in your store; the information in chat comes from there.
  • Knowledge base: Add your size chart, shipping times, and exchange and return policies.

We cover how to fill the knowledge base and what to avoid separately in our chatbot knowledge base guide. There's an added benefit: when a question has no answer in the knowledge base, the AI doesn't make one up; it hands the conversation to an agent.

If your products are on Trendyol, Hepsiburada (two of Turkey's largest online marketplaces) or Shopify, you can connect your store through marketplace integration. This integration isn't included in the Starter plan but comes with the other plans.

What determines the cost?

There are three items, and each needs to be considered separately.

  1. Chatopya plan: Monthly list prices are Starter 1,499 TL, Pro 3,999 TL and Business 6,999 TL; annual billing gets a 20% discount, and Enterprise is priced on request. Prices may change; see the pricing page for current plan details.
  2. AI credits: AI Bot usage is credit-based. The Pro plan includes 500 AI credits and the Business plan 1,000; you can buy additional credit packs if you need more.
  3. Meta message fees: Since October 1, 2026, Meta charges per message for replies sent within the 24-hour window after a customer writes. Each business number gets 1,000 free service messages per month; after that, they're charged. This fee is passed on at Meta's unit price, and Chatopya adds no commission.

The "customer wrote first, so replies are free" era is over Product recommendation is by nature a conversation of several messages. Keeping questions short and clear and cutting unnecessary in-between messages helps both your customer and your budget. The free entry point window that opens when a customer arrives from a click-to-WhatsApp ad remains free.

Frequently asked questions

Will the AI recommend products that are out of stock?

Recommendations are based on stock data updated at regular intervals from your connected store. Because updates aren't real time, we recommend confirming stock before the order for items down to their last few units.

Share this article WhatsApp LinkedIn X Facebook
Related Chatopya solution Explore marketplace integrations

Let your sales assistant get to know your catalog

Let's look together in a free demo at how AI product recommendations work with your products, stock and pricing data.

View Plans