People no longer have the same reservations about talking to machines as they might have just a few years ago. Already, 65% of smartphone owners report using voice assistants on their phones. Advances in machine learning and natural language processing have allowed chatbots to become more practical and utilitarian than ever.

Chatbots are computer programs designed to communicate with humans via voice or text. Early applications of chatbots include Map applications that give verbal directions while driving and can understand users’ verbal request for directions as well.

AI assistants like Siri, Bixby, and Alexa are also examples of chatbots that can understand to some degree what humans are saying or asking of them and respond in kind. Now we’re beginning to see more task-specific chatbots designed to work on a single website or within a single application to connect customers to brands. But as brands rush out to develop chatbots because it’s the trendy thing to do, they aren’t taking care to really integrate their chatbot into their business model and as a result, their chatbot isn’t as effective as it could be. A chatbot can potentially add a lot to your business, but only if it’s integrated properly.

What will your bot do for you?

You shouldn’t be developing a chatbot just to have one. Before you invest one dollar into developing a chatbot, you need to have clear goals in mind for it. What challenge will the chatbot solve for your business? Which performance indicators do you hope to achieve with your chatbot?

How will it work?

Chatbots are complex computer programs. They have to have access to lots of data to make them accurate. Where will that data come from. Do you already have stored transcripts of customer service interactions that you can feed them. Will you be able to grant your chatbot access to databases of information that it will need to succeed? How will you use Machine Learning and Natural Language Processing to make your chatbot better at conversing in a way that seems natural? What will your bot do when it encounters a question or statement that it doesn’t know how to respond to.

Where will it fit in?

How will your bot be integrated into your existing staff. If your bot will be functioning primarily in a customer service role, how will support tickets be divided across customer support staff and your chatbot? Will your chatbot be able to hand customers over to live agents if it gets stuck? If your chatbot is designed to aid customers in the online shopping experience, will it be able to get through the entire processing of narrowing in on the right product all the way through the checkout or will your chatbot need to be integrated with a sales team that can take over at some point in the interaction? Perhaps you plan to create a chatbot to be used for employees rather than customers. How can your chatbot bridge the gap between employees and IT or HR?

Having answers to these questions up front will allow you to be wiser as you invest in a chatbot that will be able to help your business rise to the next level.