Chatbots: Love them or hate them?

The chatbots, also known as talkbots or chatterbots, are applications in which the aim is to create the illusion of the user being speaking to a person when in fact the conversation partner is a machine. The most basic agents carry on the interaction through a chat-like textual interface, though others use sofisticated systems that try to understand spoken speech and, in the same way, create artificial speech. Some of them do a mix of both. Everybody knows at least one: Siri, that operates in many iPhones out there.

chatbots like siri


Many of these chatbots claim to be powered by artificial intelligence and a distinction needs to be made here. Some of them are and some of them are not. It can be said that in this context there are, roughly, two families of chatbots: intelligent and scripted. Intelligent chatbots use specific techniques to learn and remember what the user said or the current state of the conversation, while the scripted ones just look for keywords on the user input in order to create an answer that looks natural in the context of the conversation.

The second factor that can influence the success rate of the conversation is the access that the machine has to information, what is known as the body of knowledge. This information is used in order to compose the answers to the questions or carry on answering the user.

The last element that can change the usability of the system is the interface itself. While this seems to be the easiest to manipulate, it is certainly difficult to fine tune.

Of course, the quality of the output and the interaction varies dramatically between the different chatbots, as all these factors have different considerations on design and execution.


Certainly, most big actors of the online world have been interested on the possibilities of these applications, most of them shaping the bots as virtual assistants. Google bought in september 2016 in order to get a better positioning on the market. Slack, company that aims to be the chosen communications workspace, also exposes an API that allows the users to create conversational bots that can react to commands. Apple is getting a slower approach with the chatbot integration on iMessage through iMessage Apps and Facebook is also incorporating this technologies on its Messenger. Microsoft had a bitter experience with its Tay experiment of a Twitter bot, though they are still pursuing better results.

On the other side, many of the users seem to be dissapointed by the performance of this integrations. Many media outlets have published articles that ditch this new ways of interacting with companies, referring to the bots as ”frustrating, dull conversationalists and unnatural” or ”awful and terrible”.


In the marketing world, chatbots are primarily used as sales agents or as substitutions of first stage client service. There are many services that can be integrated in webs. This allows companies to build chat-like text interfaces that analyze the user input and answer accordingly without writing code. Chatfuel is one of the alternatives that can be used on facebook pages and Smooch allows to integrate several services on a conversational style.


Of course, this depends on both the nature of the clients, or users of the chatbot and the quality of the chatbot itself. It can be used as a first approach for getting prospective clients on your side, or even for easing typical sales interactions (”I want to order a pizza with pepperoni and bacon”). On the other hand, there are still challenges to solve, specially regarding fluidity of the conversation, naturality and size of the knowledge base. But hey, the future may be tomorrow, and maybe on it we all interact with chatbots on a daily basis.