We built Stur because African sellers were losing money in places they could not see. A customer adds one product, pays, leaves, and never hears about the three other items that would have doubled the order. Multiply that across 50 chats a week and you understand why average order value matters more than almost any other number on your dashboard.
Today we are rolling out Stur AI Recommends, a feature that watches every conversation in your storefront, understands what your customer is buying, and quietly slides in the perfect upsell or bundle in real time. No new app. No new dashboard to learn. It runs inside the same WhatsApp, Instagram, and Facebook chats you already use.
Here is what it does, how it works, and why the sellers in our beta saw their average order value lift in the first week.
What AI Recommends Actually Means
Plenty of platforms claim to use AI. Most of what they sell is a static recommendation widget, a frequently-bought-together block copied from a Western web store and stuck on a chat. That is not how Nigerian, Kenyan, or Ghanaian buyers shop. They ask questions. They negotiate. They share product photos in groups before they commit.
Stur AI Recommends works the way your buyers actually talk. When a customer in a WhatsApp chat says I want the lace dress, Stur instantly confirms the dress is in stock, looks at what other customers who bought that dress also bought, and suggests one complementary item, in a single short message, in the seller's own voice.
It does not spam. It does not break the flow. It just adds one line, at the moment the buyer is most likely to say yes.
Why We Built This for African Sellers First
Most chat-commerce AI is trained on Amazon-style behaviour: catalog, search bar, checkout. African chat commerce is different. It is conversational, social, and trust-driven. A buyer might ask is this available three times before paying. She might want to see a video. She might want a price negotiation. She might bring her sister into the chat.
We trained Stur on the way these conversations actually happen, on the language sellers in Lagos, Abuja, Nairobi, and Accra already use. That is why the recommendations land. They do not sound like a robot reading from a script. They sound like the kind of suggestion the seller's smartest assistant would make.
How AI Recommends Works Under the Hood
You do not need to understand the technology, but here is the short version, because some of you will ask.
The item the customer is asking about. Colour, size, price band, category.
The buyer history. What customers who bought this item also bought, both in your store and across similar African storefronts.
The conversation context. Is the customer asking for a gift? For a bundle? For quick gift wrap?
It combines those signals and generates a single recommendation, written in the seller's chosen tone, formal, casual, or pidgin. The seller can edit, approve, or auto-send.
You stay in control. The AI just makes sure you do not miss the moment.
Three Ways Sellers Are Already Using It
In our private beta, we watched merchants use AI Recommends in ways we did not even predict. Three patterns stood out.
The Skincare Sequencer
A skincare brand in Lekki uses AI Recommends to upsell a toner whenever a customer buys a cleanser. The AI sends a single line: most of my customers pair this with the rose toner, should I add a bottle? The brand reports a meaningful lift in basket size, and almost zero complaints, because the message is short and helpful.
The Fashion Bundle
A fashion seller in Aba uses AI Recommends to bundle accessories with every dress. When a customer adds a dress, Stur suggests a matching head wrap at a discounted bundle price. Customers who used to buy one item now often buy two.
The Reorder Nudge
A food brand in Nairobi uses AI Recommends not for first orders but for follow-ups. Thirty days after a customer buys a 500g jar of chilli oil, Stur sends a quick time for a refill message, with a one-tap reorder button. Repeat-customer revenue is the cheapest revenue in the world, and AI Recommends just makes sure no one slips through.
Setting It Up Takes Two Minutes
If you already have a Stur store, AI Recommends turned on automatically this week. Open your dashboard, click AI Settings, and choose how aggressive you want the suggestions to be, gentle, balanced, or assertive. That is it.
If you are not on Stur yet, you can get a full storefront live in about five minutes. You connect your WhatsApp Business number, upload your products with a photo and a price, connect Paystack or Flutterwave, and share your storefront link. The AI handles catalog browsing, conversational checkout, payment confirmation, order tracking, and now intelligent upsells.
A Note on Trust
When sellers first hear AI is going to talk to my customers, the first reaction is usually fear. What if it says the wrong thing? What if it offends someone?
Stur lets you set the tone, approve every new recommendation type, and switch the feature off in one tap. The AI is opinionated, but it does not act outside the boundaries you set. You can pause, edit, or fully manage every suggestion until you trust it. Most sellers stop checking after the first week.
An AI storefront should respect your relationship with your customers. We built AI Recommends to feel like a smart team member, not a chatbot.
What is Next
This is the first of three AI features shipping this quarter. Coming next: AI replies in your customer's language of choice, Yoruba, Hausa, Igbo, Pidgin, Swahili, French, and an AI-powered restock alert that watches your sales velocity and tells you what to reorder.
Try AI Recommends Free
If you already run a Stur store, log in and start using AI Recommends today, it is on by default. If you do not, set up your free Stur store in five minutes at stur.africa and start selling smarter with the first AI-native storefront built for Africa.