Walk through any market in Nairobi and watch how people buy things. They don't browse in silence. They talk to the seller. "What's this made of?" "Do you have it in a bigger size?" "Can you do 500 instead of 600?" The conversation IS the shopping experience. Websites stripped that away -- they turned shopping into a solitary scroll through product pages with a cart and a checkout button. For millions of customers in Africa, that never felt natural.
Conversational commerce brings the market stall experience online. Instead of navigating a website, the customer messages a business on WhatsApp: "Do you have running shoes in size 43?" The AI agent checks inventory, shows options, answers questions, and completes the sale -- all in a single chat thread. No website. No app download. No account creation. Just a conversation that ends with a purchase.
Why Conversational Commerce Is Growing in Africa
Three structural factors make Africa uniquely suited for chat-based commerce.
Mobile-first internet. Over 80% of internet users in Sub-Saharan Africa access the web primarily through smartphones (GSMA, 2025). Many of these devices have limited storage, making app downloads a barrier. WhatsApp is already installed on virtually every smartphone. There's no additional friction to start a conversation.
Low trust in online transactions. E-commerce adoption in Kenya is growing but many consumers still hesitate to enter payment details on unfamiliar websites. WhatsApp feels personal and familiar. Talking to a business on WhatsApp feels like talking to a person. That perceived human connection reduces purchase anxiety.
Relationship-driven buying. African commerce is fundamentally relationship-oriented. Customers prefer to buy from people they've interacted with, not faceless checkout pages. A WhatsApp conversation creates rapport. The customer can ask questions, negotiate, and feel confident before committing. That rapport translates to higher conversion rates and repeat purchases.
What Conversational Commerce Looks Like in Practice
Discovery
Traditional e-commerce: Customer searches Google, lands on your website, browses 15 product pages, opens 6 tabs, compares options.
Conversational commerce: Customer messages you: "I need a birthday gift for my wife, budget around 5K, she likes skincare." The AI agent asks a few questions about preferences and suggests three curated options with photos and prices. The customer picks one.
The conversation replaces the entire browse-and-filter workflow. Instead of the customer doing the work, the agent does it for them.
Consideration
Traditional: Customer reads product descriptions, checks reviews on three different sites, compares specifications, wonders if the photo matches reality.
Conversational: "Is this moisturizer good for oily skin?" "Can I see it on darker skin tones?" "How big is the 200ml bottle actually -- like, compared to a phone?" The agent answers with specifics from the product knowledge base, sends additional photos, and addresses concerns in real-time. Questions that would cause cart abandonment on a website get resolved in seconds.
Purchase
Traditional: Add to cart, create account, enter shipping address, enter payment details, confirm. 5-step funnel with a 70%+ abandonment rate at checkout (Baymard Institute, 2025).
Conversational: "I'll take it." Agent confirms the order, confirms the delivery address from memory or asks for it, and handles payment -- all in the same chat. No account creation. No login. No form fields. The customer never leaves WhatsApp.
Post-Purchase
Traditional: Automated email. Checked once. Forgotten.
Conversational: "Your order shipped today! Here's the tracking number. Expected delivery is Thursday. I'll message you when it's out for delivery." Thursday: "Your package is on the way! The driver is about 30 minutes out." Friday: "Hi Sarah, how's the moisturizer working out? If you love it, the matching serum is 20% off this week."
The ongoing WhatsApp thread becomes the entire customer relationship.
The Economics of Conversational Commerce
Businesses running conversational commerce on WhatsApp consistently see better unit economics than traditional e-commerce:
Higher conversion rates. Website visitors convert at 1-3% on average. WhatsApp conversations convert at 15-35% because every person who messages has already expressed intent. They're not just browsing -- they're talking to you.
Higher average order value. Conversational agents can upsell and cross-sell naturally: "That dress would look amazing with the belt we just got in. Want to see it?" This is the equivalent of the shop assistant suggesting accessories -- except it happens automatically.
Lower return rates. When customers can ask detailed questions before buying ("Will this fit if I'm between sizes?" "Is the color more navy or royal blue?"), they make better-informed purchases. One fashion retailer in Nairobi saw their return rate drop from 18% to 6% after moving to conversational sales.
Stronger retention. Customers who buy through conversation come back. The WhatsApp thread is a persistent connection. They don't need to remember your website URL or search for you again. They just scroll to your chat and message. Repeat purchase rates are typically 2-3x higher for conversational customers versus website-only customers.
Building Your Conversational Storefront
You don't need to rebuild your business from scratch. Here's how to layer conversational commerce onto what you already have.
Start With Your Top 20 Products
You don't need your entire catalog in the system on day one. Upload your 20 best-selling products with descriptions, prices, photos, and sizing information. As conversations reveal what customers ask about, expand.
Write a Sales-Oriented Agent Prompt
Your agent prompt for commerce is different from a support prompt. It should:
- Greet warmly and ask what the customer is looking for.
- Suggest products based on stated preferences, budget, and occasion.
- Answer product questions from the knowledge base.
- Handle objections ("Is this authentic?" "What if it doesn't fit?").
- Close the sale with a clear next step.
- Suggest complementary products after the initial selection.
Example: "You are the shopping assistant for Nuru Skincare. Help customers find the right products for their skin type and concerns. Ask about their skin type (oily, dry, combination, sensitive) and primary concern (acne, aging, hydration, dark spots) before recommending. Always suggest a routine (cleanser + treatment + moisturizer) rather than individual products. Be warm, knowledgeable, and honest -- if a product isn't right for their skin type, say so and recommend an alternative."
Connect Payment
When the customer is ready to buy, the agent can confirm the order and direct them through your payment flow. For businesses with M-Pesa integration on KasiLabs (coming soon), the entire transaction will happen inside the chat. For now, you can provide payment instructions and confirm receipt manually.
Track and Iterate
Pay attention to which products customers ask about most, what questions come up repeatedly, and where conversations stall. Each insight is a signal to improve your product descriptions, add to your knowledge base, or adjust your agent's approach.
Who's Already Doing This
Fashion and apparel. An online clothing brand in Lagos switched from website-only sales to WhatsApp-first and saw their monthly revenue increase 40% while reducing customer acquisition costs. The key was the AI agent's ability to help customers navigate sizing across different brands.
Beauty and skincare. A skincare company in Nairobi uses their WhatsApp agent as a skin consultation tool. Customers describe their skin concerns, and the agent recommends a personalized routine. Their average order value went from one product (KES 1,200) to three products (KES 3,400) because the agent naturally recommends complete routines.
Food and grocery. A specialty food importer uses WhatsApp for weekly ordering. Regular customers message "the usual" and the agent, remembering their previous orders, confirms and arranges delivery. Reorder friction dropped to near zero.
Home goods and furniture. A furniture store sends photos and dimensions via WhatsApp when customers describe what they're looking for. Customers can measure their space, send photos of their room, and the agent suggests pieces that fit. The conversation closes deals that would have been lost to indecision on a website.
The Future Is Already Here
Conversational commerce isn't a prediction. It's happening today across Africa, Southeast Asia, and Latin America -- markets where mobile messaging is the dominant digital channel. Meta reported that over 600 million conversations happen between businesses and consumers on WhatsApp every day globally, and that number is growing 40% year-over-year.
For businesses in Kenya, the opportunity is clear. Your customers are already on WhatsApp. They already message businesses. The question is whether they're messaging you -- and whether you're responding fast enough, helpfully enough, and at scale.