In Colombia, many businesses serve their customers through WhatsApp. It is a close, fast channel, but when messages pile up, the team gets overwhelmed. A chatbot can help. A chatbot is a program that talks with people by message and answers based on rules or with the help of artificial intelligence. But it isn't right for everything, and it's worth knowing that before you start.
What can be automated
Chatbots work best on repetitive tasks, with answers that don't change from one conversation to the next:
- Frequently asked questions. Hours, location, requirements, payment methods, status of a request.
- First contact. Greet the person, understand what they need and ask for basic details so someone on the team can pick up from there.
- Scheduling. Offer available times, confirm an appointment and send reminders.
- Information lookup. If the chatbot connects to your system, it can report the status of an order or a request.
- Request sorting. Send each message to the right team or person.
- Data collection. Conversational forms to record orders, complaints or sign-ups.
- Customer notices. Informational messages the person has agreed to receive.
What is better left to a person
- Sensitive or emotionally charged cases. A serious complaint or an upset customer needs someone who listens closely.
- Negotiations and decisions. Exceptions, special agreements or situations your rules don't cover.
- Questions the chatbot doesn't understand. If the system doesn't know, it should say so and hand the conversation to a person, not make something up.
- Legal, medical or financial topics that call for professional judgment. The chatbot can point people to the right service, but the decision belongs to a professional.
The practical rule: automate what repeats and leave to a person what requires judgment. A good chatbot always offers a way out to a human.
Two types of chatbot
Some chatbots are rule-based: they follow a menu and a tree of options defined in advance. They are predictable and work well when questions are few and clear.
Others use artificial intelligence to interpret what the person writes in their own words. They understand better, but they need clear limits: which sources they can answer from and what to do when they don't know. Many projects combine both.
Before building, define three things
- A specific use case. Start with one task, such as booking appointments or answering the most common questions. Expanding later is easier than fixing a chatbot that tried to do everything.
- A responsible person. Someone on the team must review conversations, correct answers and handle what the chatbot passes on.
- Data handling. Conversations contain personal data. Define what is stored, who sees it and how people are informed, in line with Law 1581 of 2012.
How to tell if it works
You don't need a long list of metrics. Check whether people complete what they came to do, how many conversations move to the human team and which questions the chatbot couldn't resolve. That last group is your improvement list.
What we do
At ParqueSoft Crea we build chatbots and assistants connected to your company's systems, and we support you through implementation and ongoing support. Before building, we talk with you to choose the use case that makes sense to start with.
If you want the full picture, read How to implement AI in your business starting with the use case, not the technology.
Facing a similar challenge? Get a quote
