Chatbot With Product Recommendations: Three Ways to Do It With Your Own Catalogue, and a Scorecard to Pick the One That Fits (Free Method Scorecard)
Chatbot with product recommendations can mean three different things: an AI that matches questions to your product notes, a set of guided messages with fixed suggestions, or a person choosing from the inbox. Each suits a different shop. LiveChatBot does not read your product feed or sync with Shopify or WooCommerce catalogues, so the recommendations come from what you teach it. This guide compares the three methods and gives you a scorecard to weigh them for your own shop.
⚡ In short
- LiveChatBot recommends only what you write down. It does not pull your catalogue or stock levels.
- Method 1: the AI Agent, taught with product notes. Method 2: Flow messages and saved replies. Method 3: a person in the inbox.
- Small catalogues suit fixed suggestions; large or changing ones suit AI plus a person.
- Use the scorecard to weigh the methods by what matters in your shop.

The truth
What a chatbot with product recommendations really takes
Good recommendations need two things: knowledge of the buyer’s need and knowledge of your products. A bot can ask about the need. The product knowledge has to come from you.
What LiveChatBot cannot do
It cannot see live stock, prices or variants from Shopify or WooCommerce. There is no Shopify app; Shopify is paste-the-code, and WooCommerce uses the WordPress plugin for the chat only. So the bot must not promise availability.
What it can do
It can ask what the buyer needs, match the need to the product notes you wrote, and send the buyer to the right page. Or it can pass the question to your team with the context of the page the visitor is on.
Why the method matters
Pick the wrong method for a chatbot with product recommendations and you either spend hours writing notes nobody reads, or let a bot guess. For a closer look at writing notes, see the AI sales assistant guide.
Method one
Method one: the AI Agent with product notes
With the AI Agent, you write short notes in Knowledge and the bot answers from them.
What a good note looks like
“Everyday running shoe, cushioned, suits flat feet, sizes 6 to 11, about ₹3,200, page: /shoes/everyday.” One product or product group per note, with the buyer it suits and the page link.
Rules for behaviour
In Rules, write: “Recommend up to three products. Always link the page. Never promise stock or delivery dates. If unsure, offer to ask the team.” Rules work while the website has AI replies, which is premium; facts go in Knowledge.
Best for
Shops with a few dozen products or product groups, where the questions repeat: which one for my need, what is the difference between X and Y.
Watch out
Update the notes when prices or ranges change. Stale notes produce confident wrong answers. See how to train your AI chatbot.
Method two
Method two: guided messages and saved replies
If you want full control, skip AI and use chat flows with saved replies.
How it works
A proactive message on a category page asks: “Looking for something specific? Tell us your budget and use.” The visitor answers in chat, and the free keyword bot replies with a saved reply that matches words like “gift” or “under 2000”. Each reply has two or three suggestions and links.
Best for
Tiny catalogues or seasonal ranges, where you can predict the questions. It costs nothing in AI allowance.
The limit
A keyword bot matches words, not meaning. It misses new phrasing, so check the chat history weekly and add the missed words.
Method three: a person
When the product is complex, a human in the inbox with the Android app gives the best advice. Use saved replies with {name} to speed it up, and a Flow message to invite the question.
Try it
Score the three methods for your shop
Set how much each factor matters (1 to 5), then score each method. The highest total is your best fit.
How quickly can you launch?
Importance Score Score Score
Hundreds of products or many variants.
Importance Score Score Score
No wrong or invented claims.
Importance Score Score Score
Little weekly maintenance.
Importance Score Score Score
Answers outside your working hours.
Importance Score Score Score
Examples
What a chatbot with product recommendations should say in chat
Here is an illustrative exchange for a made-up shoe shop.
Visitor: I need shoes for walking on a tiled floor all day.
Bot: For long days on hard floors, our Everyday Cushion (about ₹3,200) and the Comfort Stride (about ₹2,800) are the usual picks. Everyday Cushion has thicker soles. Here are both pages. Want me to ask the team about your size?
Notice what the bot did. It gave two options, not twelve. It named a reason. It linked the pages. It did not claim stock. And it offered a person.
Write the notes in an afternoon
Start with your ten best sellers and your five most common buyer needs. For each need, write one line: who it is for, the two products that fit, and the page link. That is fifteen short notes, and it covers most chats. Add a note every time the team answers a new question twice. After a month you have a knowledge base shaped by real customers, not guesses.
Measure it
Open the inbox weekly and read ten chats. Count how many ended with a click to a product page, a question to the team, or silence. Silence means the suggestion missed. Fix the note, not the bot.
Three habits to copy
- Two or three suggestions at most. More options mean fewer decisions.
- A reason for each. "Thicker soles" beats "popular".
- A next step. A link, a question or a hand-over.
Pitfalls
Mistakes that break a chatbot with product recommendations
Four mistakes cause most of the damage.
- Promising stock. The bot cannot see inventory. Write that in Rules and say "check the page for availability".
- Quoting old prices. Put a review date in each note and check monthly.
- Recommending everything. If the bot lists ten products, it has not recommended any.
- No escape route. Always offer the team. Our ecommerce live chat guide covers hand-over habits.
Also remember the badge: "Powered by LiveChatBot" stays on every plan, and there is no WhatsApp inbox, so keep sales conversations in the chat inbox or move them by hand.
Frequently asked questions
Does LiveChatBot read my product catalogue?
No. It does not sync with Shopify or WooCommerce catalogues. You write the product notes it uses.
Can the bot say if something is in stock?
No. It cannot see inventory. Tell it in Rules to point to the product page for availability.
Do I need the AI Agent for product recommendations?
No. Flow messages and saved replies from the free keyword bot work for small catalogues. AI replies are premium.
How many products should it suggest?
Two or three, each with a short reason and a link.
Is there a free way to start?
Yes. The free plan covers 2 websites and 100 free visitors per website each month, with no per-seat charges.
Keep reading
Related guides
Chatbot With Meeting Booking for B2B and Service Teams: Qualify the Request, Share the Meeting Link and Cover Your Week (Free Coverage Grid)
Chatbot with Checkout Assistance: What It Can Honestly Do at the Pay Step, What It Cannot, and a 10-Point Readiness Checklist
Chatbot With Appointment Scheduling: Offer Two Time Slots, Confirm by Hand and Keep Rescheduling Painless (Free Appointment Life Timeline)
Recommend with your own words
Flows, saved replies and the AI Agent, on a free plan for two websites. Start free on livechatbot.in.