A Lagos building materials trader calls the same three suppliers every morning before opening her shop. Cement price. Steel rod price. Sand per ton. The call takes 40 minutes. By noon, two of the prices have changed.
She is not looking for an AI tool. She is looking for one thing: the right price before she quotes a customer. She has been solving this problem the same way for four years.
The information she needs exists. It lives in WhatsApp group threads, Facebook market pages, informal SMS chains, and daily supplier calls. None of it is in one place. None of it updates itself. The overhead is not large. It is constant.
Urban micro-business operators in Lagos and Nairobi spend 2 to 4 hours per week on manual market information gathering. The workflow is not a technology gap - it is an information distribution gap. The customer profile is specific: monthly revenue between KES 30,000 and KES 200,000, or NGN 150,000 and NGN 800,000. Business runs primarily through WhatsApp. Not app-download-comfortable. Has used at least one WhatsApp bot before - a bank service, a telco, an order confirmation. Knows how to interact with a number that replies automatically.
This is not the African professional who uses Claude to draft emails. That user exists. She is not the ICP for this product.
The ICP for this product does not need an AI tool. She needs a supplier call she does not have to make herself.
Three things changed in 2024 and 2025 that make this product viable today.
Claude and Gemini both released WhatsApp-compatible API integrations in 2025. WhatsApp Business API pricing fell relative to 2022 rates. The BSP markup tier for low-volume senders is now workable at the price point this market can pay. Commodity price data from the Kenya Bureau of Standards and Lagos trade associations became more programmatically accessible than in 2023.
The architecture for this product exists. It did not exist at reasonable cost in 2023. The reason it has not been built is not technology. It is distribution.
The entry angle for an AI automation product in the African SME market is not an AI agent. It is a daily price summary delivered over WhatsApp. One specific vertical: food commodities in Nairobi, or building materials in Lagos. The subscriber does not interact with it. The number sends. The subscriber reads. Free for 30 days. KES 200 per month or NGN 1,000 per month after that.
The AI layer does not exist on launch day. It arrives in month three, after the distribution channel is proven and the price data validated. A product that starts as a daily WhatsApp newsletter and evolves into a conversational price tool is not simpler. It is better-distributed.
The market friction is not technical. Trust in information accuracy is the primary barrier. One wrong price loses the subscriber permanently. A market vendor who quotes a cement price based on an incorrect WhatsApp summary loses the customer and the credibility. The cost of one error is not one churn event. It is the entire referral network that subscriber carried.
Cold advertising does not reach this ICP. The acquisition path that works is through existing micro-business WhatsApp groups: join as a member, add value before promoting. Partnerships with SACCO groups and market associations provide the access and credibility that cold acquisition cannot replicate. A single partnership with a Nairobi food market association unlocks a subscriber base that cold acquisition cannot reach in 12 months.
The constraint is the product. An AI price tool that requires an app download or a stable data connection has not built for this market. It has built for Sandton. The product that wins delivers the right price over the channel the trader already trusts. It arrives before the first customer does.
