What Is a Chatbot and How Do Businesses Use Them?

A chatbot is software that simulates a conversation with a human user, typically through text. It receives a message, processes it, and sends back a response, either by following pre-written rules or by using AI to interpret the message and generate a reply.

The two types of chatbots, clearly separated

Not all chatbots work the same way, and confusing the two types leads to wrong expectations.

Rule-based chatbots follow a script. They present options ("Press 1 for balance, 2 for mini statement"), match keywords, and deliver pre-written responses. The WhatsApp auto-reply you get from a Kenyan business at midnight saying "We are currently closed, we will respond at 8 a.m." is a rule-based chatbot. It does not understand your message. It detected that a message arrived outside business hours and sent a fixed reply.

AI-powered chatbots use language models to interpret messages and generate responses. They can handle questions they were not specifically programmed for, understand variations in phrasing, and hold multi-turn conversations. When you ask a banking app "How do I reverse an M-Pesa transaction?" and get a helpful paragraph rather than a menu of numbered options, that is likely an AI chatbot processing your question and constructing a response.

The distinction matters for business decisions. Rule-based chatbots are cheaper, faster to set up, and perfectly adequate for businesses with predictable, repetitive inquiries. AI chatbots handle more complex conversations but cost more and require more careful setup and monitoring.

How chatbots actually work behind the scenes

A rule-based chatbot runs on decision trees. When a message arrives, the system checks it against a list of conditions. If the message contains "price," send the price list. If it contains "location," send the address. If nothing matches, send a fallback message: "Sorry, I did not understand. Please choose from these options."

An AI chatbot processes the message differently. It sends the user's text to a language model (similar to the technology behind ChatGPT or Claude). The model interprets the intent, considers the conversation history, and generates a response. The chatbot platform then delivers that response to the user.

Many modern chatbots are hybrids. They use AI for understanding the message and determining intent, then pull pre-approved responses for common questions. This keeps the AI's language understanding while ensuring responses stay accurate and on-brand.

How Kenyan businesses are using chatbots today

Chatbot adoption in Kenya has grown fastest on WhatsApp, which makes sense given that WhatsApp is the primary messaging platform for most Kenyans.

Customer service for banks and telcos. Equity Bank, KCB, and Safaricom all operate chatbots that handle common queries: checking balances, finding branch locations, understanding product features, and resolving basic issues. These bots deflect a significant volume of calls from human agents, reducing wait times for everyone.

Order taking for food and retail businesses. Restaurants and small retailers in Nairobi and Mombasa use WhatsApp chatbots to receive orders. A customer sends "I want 2 large pizzas," the bot confirms the order, provides the M-Pesa Paybill number, and confirms payment. The kitchen gets a structured order without a phone call.

Appointment scheduling for clinics and salons. Service businesses use chatbots to let clients book time slots. The bot presents available times, confirms the booking, and sends a reminder the day before. This replaces the back-and-forth phone calls that consume staff time.

Lead qualification for agencies and consultancies. A potential client messages asking about services. The chatbot asks a few qualifying questions (budget range, timeline, project type), captures the information, and either books a call with a human or provides relevant materials. The business owner wakes up to a list of qualified leads instead of vague "I'm interested" messages.

When a chatbot makes sense (and when it does not)

A chatbot is worth building when your business repeatedly answers the same questions or performs the same conversation steps. If you find yourself or your staff typing the same responses daily, that repetition is a signal.

Good candidates for chatbots: answering FAQs, confirming business hours, providing pricing information, taking simple orders, booking appointments, collecting customer details before a sales call.

Poor candidates: complex negotiations, sensitive complaints that need empathy, situations requiring deep product expertise, or conversations where the customer's trust depends on knowing they are speaking to a human. In these cases, a chatbot can still play a role by gathering initial information and then handing off to a human, but it should not attempt to handle the full conversation.

The worst chatbot experience is one that pretends to be more capable than it is. A bot that loops endlessly without solving the problem or connecting the customer to a person destroys trust faster than having no bot at all.

How to get started with a chatbot in Kenya

The simplest starting point for most Kenyan businesses is WhatsApp Business App's built-in features. You can set up auto-replies, away messages, and quick responses without any technical skills or costs.

For more sophisticated chatbots, platforms like Respond.io, WATI, and Trengo connect to the WhatsApp Business API and let you build conversation flows visually. These are paid services, but they offer far more flexibility than the basic WhatsApp Business App.

If you want an AI-powered chatbot, you can integrate language models through their APIs. This requires more technical setup but produces a bot that can handle unpredictable questions. Several Kenyan developers and agencies specialize in building these integrations for local businesses.

Regardless of the approach, start small. Map your ten most common customer questions, build responses for those, and launch. Expand based on what real users actually ask, not what you assume they will ask.

How chatbots connect to automation

A chatbot on its own answers questions. A chatbot connected to your business systems becomes an automation tool.

When a chatbot takes an order and automatically creates a record in your inventory system, that is automation. When it books an appointment and adds it to your Google Calendar, that is automation. When it collects a lead's details and adds them to your CRM, that is automation.

The chatbot is the interface. The automation is the work that happens behind it. Understanding how AI differs from simple automation helps you decide how sophisticated your chatbot needs to be.

Our free welcome module covers both concepts practically, showing how chatbots fit into broader automation workflows that Kenyan businesses can build without coding.

FAQ

How much does a chatbot cost for a small Kenyan business?

WhatsApp Business App is free and includes basic auto-reply features. Paid chatbot platforms connected to the WhatsApp Business API typically start from KES 4,999 to a monthly fee depending on the platform and message volume. Custom AI-powered chatbots built by a developer cost more upfront but can be tailored exactly to your workflow. Start with the free tier to validate that a chatbot helps before investing in paid options.

Will customers be annoyed by a chatbot?

Only if the chatbot is poorly designed. Customers get frustrated when a bot cannot answer their question and does not offer a clear path to a human. They appreciate bots that give instant answers to simple questions, especially outside business hours. The key is being transparent ("You are chatting with our automated assistant") and always providing a way to reach a real person.

Can a chatbot handle conversations in Sheng or mixed Swahili-English?

Rule-based chatbots struggle with mixed language because they rely on keyword matching. AI-powered chatbots handle it better since language models have some ability to process code-switching and informal language. However, performance varies. Testing with real messages from your actual customers is the only reliable way to know if the bot handles your audience's communication style.

Do I need the WhatsApp Business API or is the WhatsApp Business App enough?

The WhatsApp Business App is enough for businesses handling low message volumes with basic auto-replies. If you need multiple staff members responding, automated conversation flows, integration with other systems, or the ability to send bulk messages, you need the WhatsApp Business API accessed through a platform like WATI or Respond.io. The API is the foundation for building real chatbot experiences beyond simple away messages.