AI & automation
Is a WhatsApp bot worth it? A simple ROI method for Kenyan SMEs
Many Kenyan businesses already run sales on WhatsApp. The question isn't whether customers will chat; it's whether automating part of those chats earns back what it costs. Instead of quoting someone else's success numbers, here is a method you can run with your own figures in 15 minutes.
Step 1: count what is happening today
Look at the last two weeks of chats and write down:
- Chats per week that ask a question or want to buy.
- How many came in after hours (evenings, Sundays, holidays).
- How many got no reply, or a reply after more than an hour.
- The most repeated questions. Prices, location, delivery fees, opening hours and "is this available?" usually top the list.
- Average order value in KES.
- Staff time spent answering chats per day.
Step 2: estimate the two kinds of value
Recovered sales. Some slow or missed chats were sales that went to a competitor. Be conservative: assume a bot that replies instantly wins back only a small share of them.
Recovered sales per month = missed or slow chats per month × share you'd win back × your conversion rate × average order value.
Saved time. If a bot answers the repeated questions, how many staff hours does that free up?
Saved time per month = hours per day on repeat questions × share handled by the bot × working days × hourly cost.
Step 3: compare with the cost
Costs to include:
- Build: a WhatsApp bot from Captivart starts at KES 15,000; one with M-PESA checkout from KES 35,000.
- Meta's charges: Meta prices WhatsApp Business Platform messages by category. Replies inside a customer-started service window are largely free; marketing and some utility templates are charged. Check Meta's current rate card for Kenya.
- Hosting and upkeep: small monthly costs for the server and occasional updates.
If recovered sales plus saved time is clearly bigger than the monthly cost within a few months, it is worth doing.
A worked example (illustrative numbers, not a client result)
A Nairobi bakery gets 60 order chats a week, 15 of them after 7pm, and roughly a third of those get answered next morning. Average order value is KES 2,500. If an instant reply wins back even 2 of those late orders a week, that is about KES 20,000 a month in extra sales, before counting staff time saved. Swap in your own numbers; if yours are much smaller, a simple website chatbot or quick-reply templates may be enough.
When a bot is NOT worth it
- You get only a handful of chats a week.
- Every enquiry is complex and needs a human quote.
- Nobody will maintain the answers when prices change.
What a good bot does
- Answers the top 10 questions instantly, in plain English or Swahili.
- Shares the catalogue and prices, takes the order and the delivery area.
- Sends an M-PESA STK Push and confirms payment (see the M-PESA setup guide).
- Hands the chat to a person whenever it isn't sure.
Use the official platform
Unofficial "WhatsApp automation" tools that run on a normal phone number risk getting that number banned. Use the WhatsApp Business Platform (Cloud API) through Meta or a recognised provider.
Ready to run the numbers with us? See WhatsApp chatbots in Kenya, or go further with AI agents.

About the author
Kevin Milo
Founder of Captivart, a Nairobi web and AI studio. Kevin designs and builds websites, M-PESA stores, WhatsApp bots and AI agents for Kenyan businesses. More about Kevin
